Credit_Card_Default_Prediction.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/tonmoy-khanal/Credit-Card-Default-Prediction/blob/main/Credit_Card_Default_Prediction.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TOP0CAsOAmzM"
      },
      "source": [
        "# Credit Card Default Prediction\n",
        "\n",
        "#### Default Payments of Credit Card Clients in Taiwan from 2005"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#1. Exploratory Data Analysis. "
      ],
      "metadata": {
        "id": "U8WNNGcrrC7z"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "dyeObfdwAmzX"
      },
      "outputs": [],
      "source": [
        "# Import packages and modules\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DJESwhRLAmzZ",
        "outputId": "34297d18-134b-49b6-ce42-e5b843336d2c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30000, 25)"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "# Read file into dataframe and print out dataframe shapef\n",
        "df = pd.read_csv(\"/content/UCI_Credit_Card.csv\")\n",
        "df.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XNDcH-mIAmzb"
      },
      "source": [
        "## 2. Understand the Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 456
        },
        "id": "aHcICUGrAmzc",
        "outputId": "39313c72-a93a-43d4-f663-ee21594a2b2c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   ID  LIMIT_BAL  SEX  EDUCATION  MARRIAGE  AGE  PAY_0  PAY_2  PAY_3  PAY_4  \\\n",
              "0   1    20000.0    2          2         1   24      2      2     -1     -1   \n",
              "1   2   120000.0    2          2         2   26     -1      2      0      0   \n",
              "2   3    90000.0    2          2         2   34      0      0      0      0   \n",
              "3   4    50000.0    2          2         1   37      0      0      0      0   \n",
              "4   5    50000.0    1          2         1   57     -1      0     -1      0   \n",
              "5   6    50000.0    1          1         2   37      0      0      0      0   \n",
              "6   7   500000.0    1          1         2   29      0      0      0      0   \n",
              "7   8   100000.0    2          2         2   23      0     -1     -1      0   \n",
              "8   9   140000.0    2          3         1   28      0      0      2      0   \n",
              "9  10    20000.0    1          3         2   35     -2     -2     -2     -2   \n",
              "\n",
              "   ...  BILL_AMT4  BILL_AMT5  BILL_AMT6  PAY_AMT1  PAY_AMT2  PAY_AMT3  \\\n",
              "0  ...        0.0        0.0        0.0       0.0     689.0       0.0   \n",
              "1  ...     3272.0     3455.0     3261.0       0.0    1000.0    1000.0   \n",
              "2  ...    14331.0    14948.0    15549.0    1518.0    1500.0    1000.0   \n",
              "3  ...    28314.0    28959.0    29547.0    2000.0    2019.0    1200.0   \n",
              "4  ...    20940.0    19146.0    19131.0    2000.0   36681.0   10000.0   \n",
              "5  ...    19394.0    19619.0    20024.0    2500.0    1815.0     657.0   \n",
              "6  ...   542653.0   483003.0   473944.0   55000.0   40000.0   38000.0   \n",
              "7  ...      221.0     -159.0      567.0     380.0     601.0       0.0   \n",
              "8  ...    12211.0    11793.0     3719.0    3329.0       0.0     432.0   \n",
              "9  ...        0.0    13007.0    13912.0       0.0       0.0       0.0   \n",
              "\n",
              "   PAY_AMT4  PAY_AMT5  PAY_AMT6  default.payment.next.month  \n",
              "0       0.0       0.0       0.0                           1  \n",
              "1    1000.0       0.0    2000.0                           1  \n",
              "2    1000.0    1000.0    5000.0                           0  \n",
              "3    1100.0    1069.0    1000.0                           0  \n",
              "4    9000.0     689.0     679.0                           0  \n",
              "5    1000.0    1000.0     800.0                           0  \n",
              "6   20239.0   13750.0   13770.0                           0  \n",
              "7     581.0    1687.0    1542.0                           0  \n",
              "8    1000.0    1000.0    1000.0                           0  \n",
              "9   13007.0    1122.0       0.0                           0  \n",
              "\n",
              "[10 rows x 25 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-510fa7c1-5b6a-432a-bd13-e6d858ba71fd\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>ID</th>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <th>SEX</th>\n",
              "      <th>EDUCATION</th>\n",
              "      <th>MARRIAGE</th>\n",
              "      <th>AGE</th>\n",
              "      <th>PAY_0</th>\n",
              "      <th>PAY_2</th>\n",
              "      <th>PAY_3</th>\n",
              "      <th>PAY_4</th>\n",
              "      <th>...</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "      <th>PAY_AMT1</th>\n",
              "      <th>PAY_AMT2</th>\n",
              "      <th>PAY_AMT3</th>\n",
              "      <th>PAY_AMT4</th>\n",
              "      <th>PAY_AMT5</th>\n",
              "      <th>PAY_AMT6</th>\n",
              "      <th>default.payment.next.month</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>20000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>689.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>120000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>26</td>\n",
              "      <td>-1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>3272.0</td>\n",
              "      <td>3455.0</td>\n",
              "      <td>3261.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>34</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>14331.0</td>\n",
              "      <td>14948.0</td>\n",
              "      <td>15549.0</td>\n",
              "      <td>1518.0</td>\n",
              "      <td>1500.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>50000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>37</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>28314.0</td>\n",
              "      <td>28959.0</td>\n",
              "      <td>29547.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>2019.0</td>\n",
              "      <td>1200.0</td>\n",
              "      <td>1100.0</td>\n",
              "      <td>1069.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>50000.0</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>57</td>\n",
              "      <td>-1</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>20940.0</td>\n",
              "      <td>19146.0</td>\n",
              "      <td>19131.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>36681.0</td>\n",
              "      <td>10000.0</td>\n",
              "      <td>9000.0</td>\n",
              "      <td>689.0</td>\n",
              "      <td>679.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>6</td>\n",
              "      <td>50000.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>37</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>19394.0</td>\n",
              "      <td>19619.0</td>\n",
              "      <td>20024.0</td>\n",
              "      <td>2500.0</td>\n",
              "      <td>1815.0</td>\n",
              "      <td>657.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>800.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>7</td>\n",
              "      <td>500000.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>29</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>542653.0</td>\n",
              "      <td>483003.0</td>\n",
              "      <td>473944.0</td>\n",
              "      <td>55000.0</td>\n",
              "      <td>40000.0</td>\n",
              "      <td>38000.0</td>\n",
              "      <td>20239.0</td>\n",
              "      <td>13750.0</td>\n",
              "      <td>13770.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>8</td>\n",
              "      <td>100000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>23</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>221.0</td>\n",
              "      <td>-159.0</td>\n",
              "      <td>567.0</td>\n",
              "      <td>380.0</td>\n",
              "      <td>601.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>581.0</td>\n",
              "      <td>1687.0</td>\n",
              "      <td>1542.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>9</td>\n",
              "      <td>140000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>28</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>12211.0</td>\n",
              "      <td>11793.0</td>\n",
              "      <td>3719.0</td>\n",
              "      <td>3329.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>432.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>10</td>\n",
              "      <td>20000.0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>35</td>\n",
              "      <td>-2</td>\n",
              "      <td>-2</td>\n",
              "      <td>-2</td>\n",
              "      <td>-2</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>13007.0</td>\n",
              "      <td>13912.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>13007.0</td>\n",
              "      <td>1122.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10 rows × 25 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-510fa7c1-5b6a-432a-bd13-e6d858ba71fd')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-510fa7c1-5b6a-432a-bd13-e6d858ba71fd button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-510fa7c1-5b6a-432a-bd13-e6d858ba71fd');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "# Examine the first 10 rows of the data\n",
        "df.head(10)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZEeA79CiAmzd"
      },
      "source": [
        "### (1) Description of columns from dataset documentation\n",
        "There are 25 variables: <br>\n",
        "•   ID: ID of each client    <br>\n",
        "•   LIMIT_BAL: Amount of given credit in NT dollars (includes individual and family/supplementary credit  <br>\n",
        "•\tSEX: Gender (1=male, 2=female)  <br>\n",
        "•\tEDUCATION: (1=graduate school, 2=university, 3=high school, 4=others, 5=unknown, 6=unknown) <br>\n",
        "•\tMARRIAGE: Marital status (1=married, 2=single, 3=others) <br>\n",
        "•\tAGE: Age in years  <br>\n",
        "•\tPAY_0: Repayment status in September, 2005 (-1=pay duly, 1=payment delay for one month, 2=payment delay for two months, … 8=payment delay for eight months, 9=payment delay for nine months and above)  <br>\n",
        "•\tPAY_2: Repayment status in August, 2005 (scale same as above)  <br>\n",
        "•\tPAY_3: Repayment status in July, 2005 (scale same as above)  <br>\n",
        "•\tPAY_4: Repayment status in June, 2005 (scale same as above)  <br>\n",
        "•\tPAY_5: Repayment status in May, 2005 (scale same as above)  <br>\n",
        "•\tPAY_6: Repayment status in April, 2005 (scale same as above) <br>\n",
        "•\tBILL_AMT1: Amount of bill statement in September, 2005 (NT dollar) <br>\n",
        "•\tBILL_AMT2: Amount of bill statement in August, 2005 (NT dollar) <br>\n",
        "•\tBILL_AMT3: Amount of bill statement in July, 2005 (NT dollar) <br>\n",
        "•\tBILL_AMT4: Amount of bill statement in June, 2005 (NT dollar) <br>\n",
        "•\tBILL_AMT5: Amount of bill statement in May, 2005 (NT dollar) <br>\n",
        "•\tBILL_AMT6: Amount of bill statement in April, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT1: Amount of previous payment in September, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT2: Amount of previous payment in August, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT3: Amount of previous payment in July, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT4: Amount of previous payment in June, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT5: Amount of previous payment in May, 2005 (NT dollar) <br>\n",
        "•\tPAY_AMT6: Amount of previous payment in April, 2005 (NT dollar) <br>\n",
        "•\tdefault.payment.next.month: Default payment (1=yes, 0=no) <br>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2LnBjODPAmzf",
        "outputId": "2a9beb0b-2542-45f6-948e-272ebc9af665"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['ID', 'LIMIT_BAL', 'SEX', 'EDUCATION', 'MARRIAGE', 'AGE', 'PAY_0',\n",
              "       'PAY_2', 'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1', 'BILL_AMT2',\n",
              "       'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT1',\n",
              "       'PAY_AMT2', 'PAY_AMT3', 'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6',\n",
              "       'default.payment.next.month'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "# Understand the data by printing out all column names\n",
        "\n",
        "df.columns"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S_R8xQWhAmzh",
        "outputId": "cae56ff9-3400-4156-ac72-707d4b7f5096"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ID                                1.0\n",
              "LIMIT_BAL                     20000.0\n",
              "SEX                               2.0\n",
              "EDUCATION                         2.0\n",
              "MARRIAGE                          1.0\n",
              "AGE                              24.0\n",
              "PAY_0                             2.0\n",
              "PAY_2                             2.0\n",
              "PAY_3                            -1.0\n",
              "PAY_4                            -1.0\n",
              "PAY_5                            -2.0\n",
              "PAY_6                            -2.0\n",
              "BILL_AMT1                      3913.0\n",
              "BILL_AMT2                      3102.0\n",
              "BILL_AMT3                       689.0\n",
              "BILL_AMT4                         0.0\n",
              "BILL_AMT5                         0.0\n",
              "BILL_AMT6                         0.0\n",
              "PAY_AMT1                          0.0\n",
              "PAY_AMT2                        689.0\n",
              "PAY_AMT3                          0.0\n",
              "PAY_AMT4                          0.0\n",
              "PAY_AMT5                          0.0\n",
              "PAY_AMT6                          0.0\n",
              "default.payment.next.month        1.0\n",
              "Name: 0, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "# Understand the data by checking the first row\n",
        "\n",
        "df.loc[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mIZKTLbFAmzk"
      },
      "source": [
        "### (2) What do 0 and -2 mean in PAY_X columns?\n",
        "\n",
        "From dataset documentation: \n",
        "PAY_0: Repayment status in September, 2005 (-1=pay duly, 1=payment delay for one month, 2=payment delay for two months, … 8=payment delay for eight months, 9=payment delay for nine months and above). Notice code 0 and 2 are in the PAY_X columns but are not included in the data description. Also by doing the value_counts, it shows 0 is the most frequent observation. \n",
        "\n",
        "After searching online, I realized the same question was asked by others, and the answer from the dataset contributor was:\n",
        "\n",
        "#### -2: No consumption; <br>-1: Paid in full; <br>0: The use of revolving credit, meaning the payment wasn't due, which makes sense that most customers were using the revolving credit.\n",
        "\n",
        "https://www.kaggle.com/uciml/default-of-credit-card-clients-dataset/discussion/34608"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ra6Ul6enAmzl",
        "outputId": "620cfbd0-cbfd-4ab4-998a-14b31e252840"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 0    95919\n",
            "-1    34640\n",
            "-2    24415\n",
            " 2    18964\n",
            " 1     3722\n",
            " 3     1430\n",
            " 4      453\n",
            " 7      218\n",
            " 5      137\n",
            " 6       74\n",
            " 8       28\n",
            "Name: value, dtype: int64\n"
          ]
        }
      ],
      "source": [
        "# Sort the repayment status codes in PAY_X columns by melting them into 1 column\n",
        "# Notice there are 2 undocumented codes: 0,-2\n",
        "\n",
        "pay = df[['PAY_0','PAY_2','PAY_3','PAY_4','PAY_5','PAY_6']]\n",
        "pay_melt = pd.melt(pay)\n",
        "print(pay_melt['value'].value_counts())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qq3Qg9CbAmzm"
      },
      "source": [
        "### (3) What does 'default.payment.next.month' mean?\n",
        "Is it a predictor or actual data that was collected in the next month - Oct. 2005? According to the reply from the dataset contributor,it seems this is the actual data collected in the next month. \"Y: client's behavior; Y=0 then not default, Y=1 then default\". This information could be used to test the accuracy of the model that will be built."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Y5VSr0-9Amzn",
        "outputId": "f9a5deca-37d3-4af1-fea4-f08377221a0b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    0.7788\n",
              "1    0.2212\n",
              "Name: default.payment.next.month, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "# Get the proportion of customers who had default payment in the next month (Oct.2005)? \n",
        "# About 22% customers had default payment next month\n",
        "\n",
        "df['default.payment.next.month'].value_counts(normalize=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DOjc9GsVAmzo"
      },
      "source": [
        "## 3. Data Cleanning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qrsKYt8CAmzo"
      },
      "source": [
        "### (1) Check missing values, column type and column names\n",
        "There is no missing value, all columns have the correct type, and all columns are named consistently since we have renamed the last inconsistent column name to 'DEF_PAY_NMO."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v_a4kPJrAmzp",
        "outputId": "48fe80c5-582d-490a-efbc-1707d05ee353"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 30000 entries, 0 to 29999\n",
            "Data columns (total 25 columns):\n",
            " #   Column                      Non-Null Count  Dtype  \n",
            "---  ------                      --------------  -----  \n",
            " 0   ID                          30000 non-null  int64  \n",
            " 1   LIMIT_BAL                   30000 non-null  float64\n",
            " 2   SEX                         30000 non-null  int64  \n",
            " 3   EDUCATION                   30000 non-null  int64  \n",
            " 4   MARRIAGE                    30000 non-null  int64  \n",
            " 5   AGE                         30000 non-null  int64  \n",
            " 6   PAY_0                       30000 non-null  int64  \n",
            " 7   PAY_2                       30000 non-null  int64  \n",
            " 8   PAY_3                       30000 non-null  int64  \n",
            " 9   PAY_4                       30000 non-null  int64  \n",
            " 10  PAY_5                       30000 non-null  int64  \n",
            " 11  PAY_6                       30000 non-null  int64  \n",
            " 12  BILL_AMT1                   30000 non-null  float64\n",
            " 13  BILL_AMT2                   30000 non-null  float64\n",
            " 14  BILL_AMT3                   30000 non-null  float64\n",
            " 15  BILL_AMT4                   30000 non-null  float64\n",
            " 16  BILL_AMT5                   30000 non-null  float64\n",
            " 17  BILL_AMT6                   30000 non-null  float64\n",
            " 18  PAY_AMT1                    30000 non-null  float64\n",
            " 19  PAY_AMT2                    30000 non-null  float64\n",
            " 20  PAY_AMT3                    30000 non-null  float64\n",
            " 21  PAY_AMT4                    30000 non-null  float64\n",
            " 22  PAY_AMT5                    30000 non-null  float64\n",
            " 23  PAY_AMT6                    30000 non-null  float64\n",
            " 24  default.payment.next.month  30000 non-null  int64  \n",
            "dtypes: float64(13), int64(12)\n",
            "memory usage: 5.7 MB\n"
          ]
        }
      ],
      "source": [
        "# Check column names, columan data type and indentify any missing values\n",
        "df.info()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RS-rACQ6Amzq",
        "outputId": "c375bb63-62f5-48ba-9fc5-b76ae83bfbd3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['ID', 'LIMIT_BAL', 'SEX', 'EDUCATION', 'MARRIAGE', 'AGE', 'PAY_1',\n",
              "       'PAY_2', 'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1', 'BILL_AMT2',\n",
              "       'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT1',\n",
              "       'PAY_AMT2', 'PAY_AMT3', 'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6',\n",
              "       'DEF_PAY_NMO'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ],
      "source": [
        "# Change inconsistent column names \"PAY_0\" to 'PAY_1', 'default.payment.next.month'to 'DEF_PAY_NMO'\n",
        "\n",
        "df.rename(columns={'PAY_0':'PAY_1','default.payment.next.month':'DEF_PAY_NMO'},inplace = True)\n",
        "df.columns"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nF6iwQudAmzq"
      },
      "source": [
        "### (2) Check duplicate records\n",
        "There are no duplicate IDs or rows."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RIgCJnNxAmzr",
        "outputId": "f2716489-d5ac-485d-e876-3167603a66e8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "No duplicate IDs\n"
          ]
        }
      ],
      "source": [
        "# Check if there is any duplicate IDs\n",
        "condition = bool(df.duplicated(subset = 'ID').any())\n",
        "\n",
        "if condition:\n",
        "    print('There are duplicate IDs')\n",
        "else:\n",
        "    print('No duplicate IDs')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nyRua9MlAmzs"
      },
      "source": [
        "### (3) Identify outliers"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "NY_b4unrAmzs",
        "outputId": "5dac513c-e72c-4d7e-822c-6ea84a9a6e1f"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Use histogram to view the distribution of credit limit\n",
        "\n",
        "plt.hist(df[\"LIMIT_BAL\"])\n",
        "plt.xlabel('Credit_Limit in NT$', fontweight='bold')\n",
        "plt.ylabel('# of Customers', fontweight='bold')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 291
        },
        "id": "iKg1JzGRAmzt",
        "outputId": "63399cfb-d491-4a3d-b2b5-34bbf2bdbd2d"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Use boxplot to identify any outlier in credit limit\n",
        "\n",
        "df[\"LIMIT_BAL\"].plot(kind=\"box\")\n",
        "plt.xlabel('Credit limit in NT$', fontweight='bold')\n",
        "plt.ylabel('# of Customers', fontweight='bold')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        },
        "id": "QkuCPV8SAmzt",
        "outputId": "c84fb3ff-1d25-4e11-b0c5-efd10cc28cd7"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        ID  LIMIT_BAL  SEX  EDUCATION  MARRIAGE  AGE  PAY_1  PAY_2  PAY_3  \\\n",
              "2197  2198  1000000.0    2          1         1   47      0      0      0   \n",
              "\n",
              "      PAY_4  ...  BILL_AMT4  BILL_AMT5  BILL_AMT6  PAY_AMT1  PAY_AMT2  \\\n",
              "2197     -1  ...   891586.0   927171.0   961664.0   50784.0   50723.0   \n",
              "\n",
              "      PAY_AMT3  PAY_AMT4  PAY_AMT5  PAY_AMT6  DEF_PAY_NMO  \n",
              "2197  896040.0   50000.0   50000.0   50256.0            0  \n",
              "\n",
              "[1 rows x 25 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8f9baf89-ed13-4339-bebe-48a97d54a808\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>ID</th>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <th>SEX</th>\n",
              "      <th>EDUCATION</th>\n",
              "      <th>MARRIAGE</th>\n",
              "      <th>AGE</th>\n",
              "      <th>PAY_1</th>\n",
              "      <th>PAY_2</th>\n",
              "      <th>PAY_3</th>\n",
              "      <th>PAY_4</th>\n",
              "      <th>...</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "      <th>PAY_AMT1</th>\n",
              "      <th>PAY_AMT2</th>\n",
              "      <th>PAY_AMT3</th>\n",
              "      <th>PAY_AMT4</th>\n",
              "      <th>PAY_AMT5</th>\n",
              "      <th>PAY_AMT6</th>\n",
              "      <th>DEF_PAY_NMO</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2197</th>\n",
              "      <td>2198</td>\n",
              "      <td>1000000.0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>47</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>...</td>\n",
              "      <td>891586.0</td>\n",
              "      <td>927171.0</td>\n",
              "      <td>961664.0</td>\n",
              "      <td>50784.0</td>\n",
              "      <td>50723.0</td>\n",
              "      <td>896040.0</td>\n",
              "      <td>50000.0</td>\n",
              "      <td>50000.0</td>\n",
              "      <td>50256.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1 rows × 25 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8f9baf89-ed13-4339-bebe-48a97d54a808')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-8f9baf89-ed13-4339-bebe-48a97d54a808 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-8f9baf89-ed13-4339-bebe-48a97d54a808');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "# Take a close look at the outlier\n",
        "# The observations from other columns indicate that this was just a customer who has an excellent payment history\n",
        "# Therefore it should be considered as a valid data\n",
        "\n",
        "outliers = df.loc[df['LIMIT_BAL']>900000]\n",
        "outliers"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vCpr2-2tAmzu"
      },
      "source": [
        "### (4) Check numerical columns"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "FACFVgbCAmzv",
        "outputId": "0d2337f7-77c3-4faa-ba8c-bdfc124fb760"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 ID       LIMIT_BAL           SEX     EDUCATION      MARRIAGE  \\\n",
              "count  30000.000000    30000.000000  30000.000000  30000.000000  30000.000000   \n",
              "mean   15000.500000   167484.322667      1.603733      1.853133      1.551867   \n",
              "std     8660.398374   129747.661567      0.489129      0.790349      0.521970   \n",
              "min        1.000000    10000.000000      1.000000      0.000000      0.000000   \n",
              "25%     7500.750000    50000.000000      1.000000      1.000000      1.000000   \n",
              "50%    15000.500000   140000.000000      2.000000      2.000000      2.000000   \n",
              "75%    22500.250000   240000.000000      2.000000      2.000000      2.000000   \n",
              "max    30000.000000  1000000.000000      2.000000      6.000000      3.000000   \n",
              "\n",
              "                AGE         PAY_1         PAY_2         PAY_3         PAY_4  \\\n",
              "count  30000.000000  30000.000000  30000.000000  30000.000000  30000.000000   \n",
              "mean      35.485500     -0.016700     -0.133767     -0.166200     -0.220667   \n",
              "std        9.217904      1.123802      1.197186      1.196868      1.169139   \n",
              "min       21.000000     -2.000000     -2.000000     -2.000000     -2.000000   \n",
              "25%       28.000000     -1.000000     -1.000000     -1.000000     -1.000000   \n",
              "50%       34.000000      0.000000      0.000000      0.000000      0.000000   \n",
              "75%       41.000000      0.000000      0.000000      0.000000      0.000000   \n",
              "max       79.000000      8.000000      8.000000      8.000000      8.000000   \n",
              "\n",
              "       ...      BILL_AMT4      BILL_AMT5      BILL_AMT6       PAY_AMT1  \\\n",
              "count  ...   30000.000000   30000.000000   30000.000000   30000.000000   \n",
              "mean   ...   43262.948967   40311.400967   38871.760400    5663.580500   \n",
              "std    ...   64332.856134   60797.155770   59554.107537   16563.280354   \n",
              "min    ... -170000.000000  -81334.000000 -339603.000000       0.000000   \n",
              "25%    ...    2326.750000    1763.000000    1256.000000    1000.000000   \n",
              "50%    ...   19052.000000   18104.500000   17071.000000    2100.000000   \n",
              "75%    ...   54506.000000   50190.500000   49198.250000    5006.000000   \n",
              "max    ...  891586.000000  927171.000000  961664.000000  873552.000000   \n",
              "\n",
              "           PAY_AMT2      PAY_AMT3       PAY_AMT4       PAY_AMT5  \\\n",
              "count  3.000000e+04   30000.00000   30000.000000   30000.000000   \n",
              "mean   5.921163e+03    5225.68150    4826.076867    4799.387633   \n",
              "std    2.304087e+04   17606.96147   15666.159744   15278.305679   \n",
              "min    0.000000e+00       0.00000       0.000000       0.000000   \n",
              "25%    8.330000e+02     390.00000     296.000000     252.500000   \n",
              "50%    2.009000e+03    1800.00000    1500.000000    1500.000000   \n",
              "75%    5.000000e+03    4505.00000    4013.250000    4031.500000   \n",
              "max    1.684259e+06  896040.00000  621000.000000  426529.000000   \n",
              "\n",
              "            PAY_AMT6   DEF_PAY_NMO  \n",
              "count   30000.000000  30000.000000  \n",
              "mean     5215.502567      0.221200  \n",
              "std     17777.465775      0.415062  \n",
              "min         0.000000      0.000000  \n",
              "25%       117.750000      0.000000  \n",
              "50%      1500.000000      0.000000  \n",
              "75%      4000.000000      0.000000  \n",
              "max    528666.000000      1.000000  \n",
              "\n",
              "[8 rows x 25 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-0f2c0779-388f-4e81-aa5d-27a110cd0598\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>ID</th>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <th>SEX</th>\n",
              "      <th>EDUCATION</th>\n",
              "      <th>MARRIAGE</th>\n",
              "      <th>AGE</th>\n",
              "      <th>PAY_1</th>\n",
              "      <th>PAY_2</th>\n",
              "      <th>PAY_3</th>\n",
              "      <th>PAY_4</th>\n",
              "      <th>...</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "      <th>PAY_AMT1</th>\n",
              "      <th>PAY_AMT2</th>\n",
              "      <th>PAY_AMT3</th>\n",
              "      <th>PAY_AMT4</th>\n",
              "      <th>PAY_AMT5</th>\n",
              "      <th>PAY_AMT6</th>\n",
              "      <th>DEF_PAY_NMO</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>3.000000e+04</td>\n",
              "      <td>30000.00000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>15000.500000</td>\n",
              "      <td>167484.322667</td>\n",
              "      <td>1.603733</td>\n",
              "      <td>1.853133</td>\n",
              "      <td>1.551867</td>\n",
              "      <td>35.485500</td>\n",
              "      <td>-0.016700</td>\n",
              "      <td>-0.133767</td>\n",
              "      <td>-0.166200</td>\n",
              "      <td>-0.220667</td>\n",
              "      <td>...</td>\n",
              "      <td>43262.948967</td>\n",
              "      <td>40311.400967</td>\n",
              "      <td>38871.760400</td>\n",
              "      <td>5663.580500</td>\n",
              "      <td>5.921163e+03</td>\n",
              "      <td>5225.68150</td>\n",
              "      <td>4826.076867</td>\n",
              "      <td>4799.387633</td>\n",
              "      <td>5215.502567</td>\n",
              "      <td>0.221200</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>8660.398374</td>\n",
              "      <td>129747.661567</td>\n",
              "      <td>0.489129</td>\n",
              "      <td>0.790349</td>\n",
              "      <td>0.521970</td>\n",
              "      <td>9.217904</td>\n",
              "      <td>1.123802</td>\n",
              "      <td>1.197186</td>\n",
              "      <td>1.196868</td>\n",
              "      <td>1.169139</td>\n",
              "      <td>...</td>\n",
              "      <td>64332.856134</td>\n",
              "      <td>60797.155770</td>\n",
              "      <td>59554.107537</td>\n",
              "      <td>16563.280354</td>\n",
              "      <td>2.304087e+04</td>\n",
              "      <td>17606.96147</td>\n",
              "      <td>15666.159744</td>\n",
              "      <td>15278.305679</td>\n",
              "      <td>17777.465775</td>\n",
              "      <td>0.415062</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>1.000000</td>\n",
              "      <td>10000.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>21.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>-170000.000000</td>\n",
              "      <td>-81334.000000</td>\n",
              "      <td>-339603.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000e+00</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>7500.750000</td>\n",
              "      <td>50000.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>28.000000</td>\n",
              "      <td>-1.000000</td>\n",
              "      <td>-1.000000</td>\n",
              "      <td>-1.000000</td>\n",
              "      <td>-1.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>2326.750000</td>\n",
              "      <td>1763.000000</td>\n",
              "      <td>1256.000000</td>\n",
              "      <td>1000.000000</td>\n",
              "      <td>8.330000e+02</td>\n",
              "      <td>390.00000</td>\n",
              "      <td>296.000000</td>\n",
              "      <td>252.500000</td>\n",
              "      <td>117.750000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>15000.500000</td>\n",
              "      <td>140000.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>34.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>19052.000000</td>\n",
              "      <td>18104.500000</td>\n",
              "      <td>17071.000000</td>\n",
              "      <td>2100.000000</td>\n",
              "      <td>2.009000e+03</td>\n",
              "      <td>1800.00000</td>\n",
              "      <td>1500.000000</td>\n",
              "      <td>1500.000000</td>\n",
              "      <td>1500.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>22500.250000</td>\n",
              "      <td>240000.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>41.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>54506.000000</td>\n",
              "      <td>50190.500000</td>\n",
              "      <td>49198.250000</td>\n",
              "      <td>5006.000000</td>\n",
              "      <td>5.000000e+03</td>\n",
              "      <td>4505.00000</td>\n",
              "      <td>4013.250000</td>\n",
              "      <td>4031.500000</td>\n",
              "      <td>4000.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>30000.000000</td>\n",
              "      <td>1000000.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>6.000000</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>79.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>891586.000000</td>\n",
              "      <td>927171.000000</td>\n",
              "      <td>961664.000000</td>\n",
              "      <td>873552.000000</td>\n",
              "      <td>1.684259e+06</td>\n",
              "      <td>896040.00000</td>\n",
              "      <td>621000.000000</td>\n",
              "      <td>426529.000000</td>\n",
              "      <td>528666.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>8 rows × 25 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0f2c0779-388f-4e81-aa5d-27a110cd0598')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-0f2c0779-388f-4e81-aa5d-27a110cd0598 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-0f2c0779-388f-4e81-aa5d-27a110cd0598');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "# Get the statistic summary of the columns\n",
        "# No data is extremly unresonable in these columns\n",
        "\n",
        "df.describe()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "jo5sHNe5Amzv"
      },
      "outputs": [],
      "source": [
        "# Save processed csv file for part 2 - modeling\n",
        "\n",
        "df.to_csv(\"UCI_Credit_Card_Processed.csv\", index=None)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cpFaVRMqAmzv"
      },
      "source": [
        "### (5) Check categorical columns\n",
        "\n",
        "Next, we will check these 3 columns: SEX',\t'EDUCATION' and\t'MARRIAGE'.\tThe 'SEX' column looks fine, while 'EDUCATION' and 'MARRIAGE' columns have some undocumented data and 'EDUCATION' column has duplicated values in different codes."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NrinwuwUAmzw",
        "outputId": "5cb88f0f-e823-471c-a5b0-e6756107ee70"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2    18112\n",
              "1    11888\n",
              "Name: SEX, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ],
      "source": [
        "# Get SEX column's distribution. 1: male; 2: female\n",
        "# No undocumented SEX code\n",
        "\n",
        "df[\"SEX\"].value_counts()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gY7r9kwcAmzx",
        "outputId": "7db7fc78-21f6-44bc-a147-796d0926ed39"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "CategoricalDtype(categories=['F', 'M'], ordered=False)"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "# Change sex code to 'F' and 'M' and use category as datatype\n",
        "\n",
        "df[\"SEX\"]=df[\"SEX\"].map({1:'M',2:'F'}).astype('category')\n",
        "df[\"SEX\"].dtypes"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zZ6wf4NzAmzy"
      },
      "source": [
        "'EDUCATION' column: notice 5 and 6 are both recorded as 'unknown' and there is 0 which isn't explained in the dataset description. Since the amounts are so small, let's combine 0,4,5,6 to 0 which means\"other'."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BDTL1UE8Amzy",
        "outputId": "33ab70c2-df59-4098-9bbc-4d449f8682fd"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([2, 1, 3, 5, 4, 6, 0])"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "# EDUCATION: (1=graduate school, 2=university, 3=high school, 4=others, 5=unknown, 6=unknown)\n",
        "\n",
        "df[\"EDUCATION\"].unique()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2eGSkVwUAmzz",
        "outputId": "bf03e359-282f-4d68-a9b7-0dff6391e165"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2    14030\n",
              "1    10585\n",
              "3     4917\n",
              "5      280\n",
              "4      123\n",
              "6       51\n",
              "0       14\n",
              "Name: EDUCATION, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "# Get the total counts of each EDUCATION level\n",
        "\n",
        "df[\"EDUCATION\"].value_counts()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X-DOwGLdAmzz",
        "outputId": "eba9eda7-4578-480f-db0b-d1e82bdc6990"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2    14030\n",
              "1    10585\n",
              "3     4917\n",
              "0      468\n",
              "Name: EDUCATION, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ],
      "source": [
        "# Change values 4, 5, 6 to 0 and define 0 as 'others'\n",
        "# 1=graduate school, 2=university, 3=high school, 0=others\n",
        "\n",
        "df[\"EDUCATION\"] = df[\"EDUCATION\"].replace({4:0,5:0,6:0})\n",
        "df[\"EDUCATION\"].value_counts()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HIsG5tfUAmz0"
      },
      "source": [
        "'MARRIAGE' column: what does 0 mean in 'MARRIAGE'?\n",
        "Since there are only 0.18% (or 54) observations of 0, we will combine 0 and 3 in one value as 'others'."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JSvOuFMcAmz0",
        "outputId": "77f13aaf-4478-434d-efed-a62b8e5952e3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1, 2, 3, 0])"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "# From dataset description: MARRIAGE: Marital status (1=married, 2=single, 3=others), but there is also 0\n",
        "\n",
        "df[\"MARRIAGE\"].unique()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k9k5IyqQAmz1",
        "outputId": "3ec676ef-7f9c-4b51-87c2-3987841b379a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2    0.532133\n",
              "1    0.455300\n",
              "3    0.010767\n",
              "0    0.001800\n",
              "Name: MARRIAGE, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "# How many customers had \"MARRIAGE\" status as 0?\n",
        "\n",
        "df[\"MARRIAGE\"].value_counts(normalize=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9N8yDPRiAmz2",
        "outputId": "f80b71df-051d-488f-eee8-185985405b7e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2    0.532133\n",
              "1    0.455300\n",
              "3    0.012567\n",
              "Name: MARRIAGE, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "# Combine 0 and 3 by changing the value 0 into 3\n",
        "\n",
        "df[\"MARRIAGE\"] = df[\"MARRIAGE\"].replace({0:3})\n",
        "df[\"MARRIAGE\"].value_counts(normalize=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H71J2mB-Amz2"
      },
      "source": [
        "## 4. Trends, Hypotheses and Findings"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 331
        },
        "id": "iASizjSYAmz3",
        "outputId": "2e319921-5533-47d6-f082-f78a9fbcaf37"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    LIMIT_BAL SEX  EDUCATION  MARRIAGE  AGE  PAY_1  PAY_2  PAY_3  PAY_4  \\\n",
              "ID                                                                        \n",
              "1     20000.0   F          2         1   24      2      2     -1     -1   \n",
              "2    120000.0   F          2         2   26     -1      2      0      0   \n",
              "3     90000.0   F          2         2   34      0      0      0      0   \n",
              "4     50000.0   F          2         1   37      0      0      0      0   \n",
              "5     50000.0   M          2         1   57     -1      0     -1      0   \n",
              "\n",
              "    PAY_5  ...  BILL_AMT4  BILL_AMT5  BILL_AMT6  PAY_AMT1  PAY_AMT2  PAY_AMT3  \\\n",
              "ID         ...                                                                  \n",
              "1      -2  ...        0.0        0.0        0.0       0.0     689.0       0.0   \n",
              "2       0  ...     3272.0     3455.0     3261.0       0.0    1000.0    1000.0   \n",
              "3       0  ...    14331.0    14948.0    15549.0    1518.0    1500.0    1000.0   \n",
              "4       0  ...    28314.0    28959.0    29547.0    2000.0    2019.0    1200.0   \n",
              "5       0  ...    20940.0    19146.0    19131.0    2000.0   36681.0   10000.0   \n",
              "\n",
              "    PAY_AMT4  PAY_AMT5  PAY_AMT6  DEF_PAY_NMO  \n",
              "ID                                             \n",
              "1        0.0       0.0       0.0            1  \n",
              "2     1000.0       0.0    2000.0            1  \n",
              "3     1000.0    1000.0    5000.0            0  \n",
              "4     1100.0    1069.0    1000.0            0  \n",
              "5     9000.0     689.0     679.0            0  \n",
              "\n",
              "[5 rows x 24 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-203dbf43-afa1-49fa-9708-fc80d1a6473e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <th>SEX</th>\n",
              "      <th>EDUCATION</th>\n",
              "      <th>MARRIAGE</th>\n",
              "      <th>AGE</th>\n",
              "      <th>PAY_1</th>\n",
              "      <th>PAY_2</th>\n",
              "      <th>PAY_3</th>\n",
              "      <th>PAY_4</th>\n",
              "      <th>PAY_5</th>\n",
              "      <th>...</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "      <th>PAY_AMT1</th>\n",
              "      <th>PAY_AMT2</th>\n",
              "      <th>PAY_AMT3</th>\n",
              "      <th>PAY_AMT4</th>\n",
              "      <th>PAY_AMT5</th>\n",
              "      <th>PAY_AMT6</th>\n",
              "      <th>DEF_PAY_NMO</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ID</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>20000.0</td>\n",
              "      <td>F</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-2</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>689.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>120000.0</td>\n",
              "      <td>F</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>26</td>\n",
              "      <td>-1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>3272.0</td>\n",
              "      <td>3455.0</td>\n",
              "      <td>3261.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>90000.0</td>\n",
              "      <td>F</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>34</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>14331.0</td>\n",
              "      <td>14948.0</td>\n",
              "      <td>15549.0</td>\n",
              "      <td>1518.0</td>\n",
              "      <td>1500.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>50000.0</td>\n",
              "      <td>F</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>37</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>28314.0</td>\n",
              "      <td>28959.0</td>\n",
              "      <td>29547.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>2019.0</td>\n",
              "      <td>1200.0</td>\n",
              "      <td>1100.0</td>\n",
              "      <td>1069.0</td>\n",
              "      <td>1000.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>50000.0</td>\n",
              "      <td>M</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>57</td>\n",
              "      <td>-1</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>20940.0</td>\n",
              "      <td>19146.0</td>\n",
              "      <td>19131.0</td>\n",
              "      <td>2000.0</td>\n",
              "      <td>36681.0</td>\n",
              "      <td>10000.0</td>\n",
              "      <td>9000.0</td>\n",
              "      <td>689.0</td>\n",
              "      <td>679.0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 24 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-203dbf43-afa1-49fa-9708-fc80d1a6473e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-203dbf43-afa1-49fa-9708-fc80d1a6473e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-203dbf43-afa1-49fa-9708-fc80d1a6473e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ],
      "source": [
        "# Set ID as dataframe index\n",
        "\n",
        "df.set_index('ID', inplace = True)\n",
        "df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "RkStz9iqAmz3"
      },
      "outputs": [],
      "source": [
        "# Create a new column \"HAS_DEF\" to indicate customers who have at least 1 deafult payment from PAY_1 to Pay_6\n",
        "# 0 : no default ; 1: has default\n",
        "\n",
        "def_condition = (df.PAY_1 >1) | (df.PAY_2 >1) | (df.PAY_3 >1) | (df.PAY_4 >1) | (df.PAY_5 >1) | (df.PAY_6 >1)\n",
        "df.loc[def_condition, \"HAS_DEF\"] = 1\n",
        "df.loc[df.HAS_DEF.isna(), \"HAS_DEF\"] = 0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "abojqBTdAmz4"
      },
      "source": [
        "#### (1) On average, which sex group tends to have more delayed payments? Since there are more female than male in the dataset, we will use percentage within each sex group.\n",
        "\n",
        "30% male have default payment while 26% female have default payment, the difference is not significant."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "w9TzuZScAmz4"
      },
      "outputs": [],
      "source": [
        "# Define a function to plot stacked bar chart since there will be multiple stacked bars charts to plot\n",
        "\n",
        "def create_stacked_bar(column_name):\n",
        "\n",
        "    # Get the percentage of default by each group\n",
        "    default_by_group = pd.crosstab(index=df['HAS_DEF'],columns = df[column_name], normalize = 'columns')\n",
        "    \n",
        "    # Round up to 2 decimal\n",
        "    default_by_group = default_by_group.apply(lambda x: round(x,2))\n",
        "    \n",
        "    labels = default_by_group.columns\n",
        "    list1 = default_by_group.iloc[0].to_list()\n",
        "    list2 = default_by_group.iloc[1].to_list()\n",
        "    \n",
        "    list1_name = \"No default\"\n",
        "    list2_name = \"Has default\"\n",
        "    title = f\"Default by {column_name}\"\n",
        "    xlabel = column_name\n",
        "    ylabel = \"Default percentage\"\n",
        "    \n",
        "    fig, ax = plt.subplots(figsize=(10, 5))\n",
        "    bar_width = 0.5\n",
        "    \n",
        "    ax1 = ax.bar(labels,list1, bar_width, label = list1_name)\n",
        "    ax2 = ax.bar(labels,list2, bar_width, bottom = list1, label = list2_name)\n",
        "\n",
        "    ax.set_title(title, fontweight = \"bold\")\n",
        "    ax.set_xlabel(xlabel, fontweight = \"bold\")\n",
        "    ax.set_ylabel(ylabel, fontweight = \"bold\")\n",
        "    ax.legend(loc=\"best\")\n",
        "    \n",
        "    plt.xticks(list(range(len(labels))), labels,rotation=90)\n",
        "    plt.yticks(fontsize=9)\n",
        "\n",
        "    for r1, r2 in zip(ax1, ax2):\n",
        "        h1 = r1.get_height()\n",
        "        h2 = r2.get_height()\n",
        "        plt.text(r1.get_x() + r1.get_width() / 2., h1 / 2., f\"{h1:.0%}\", ha=\"center\", va=\"center\", color=\"white\", fontsize=9, fontweight=\"bold\")\n",
        "        plt.text(r2.get_x() + r2.get_width() / 2., h1 + h2 / 2., f\"{h2:.0%}\", ha=\"center\", va=\"center\", color=\"white\", fontsize=9, fontweight=\"bold\")\n",
        "\n",
        "    plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 348
        },
        "id": "gDZowySyAmz5",
        "outputId": "fd3d1382-737f-49c6-8d46-352d5aa0e863"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Plot a stacked bar chart to visualize default percentage by female and male\n",
        "\n",
        "create_stacked_bar('SEX')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Rz-DY4-zAmz5"
      },
      "source": [
        "#### (2) Did customers with higher education have less delayed payment?\n",
        "\n",
        "The data indicates customers with lower education levels default more. Customers with high school and university educational level had higher default percentages than customers with grad school education did. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "id": "9u17HxJaAmz6",
        "outputId": "d794e732-85f7-467b-d075-166f310440d3"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl8AAAFJCAYAAACl2EgGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd5hV1dmw8fuhCKiASLERRGNDFAGxo2KNvcVgjxj9MBpbYo95IxpjfKN5LTEmUWM0iiUaC2ps0aBYQgRFVLCggKKIgCii9FnfH3sPHsZh5lDOmWHm/l3Xuc5Za6+993PGc+njWmuvFSklJEmSVB5N6joASZKkxsTkS5IkqYxMviRJksrI5EuSJKmMTL4kSZLKyORLkiSpjEy+JC1RRLSJiIci4suISBFx2gq6bspfXZfinK6V562IGCSprph8SQ1QREzIE5WKiJiVl/8eEdst5aV+DBwETAeuA0at8GCBiBiaxzugFNev4b4DChLBwteD+fGuBXULImJmRLwZEddHROeC6xS2W6OgflTV7xURq0bEJRExJiLmRMT0iPh3ROxZJbZLC67522ruU+0rb1f59zyr4HrtI+K6/LcwLyI+jYh7I6L7Er7HgwX1Z+V1Q1fk319qrJrVdQCSSupRYCqwE/AD4NCIODqldG+R52+Sv/8tpfTLUgRYT0wDBheUX6+mzU1AO2AP4CdA/4jom1J6p9ibRMSqwFBgG+BrYAgwF+gLHAD8q6D5MQWfj4qIC4CZwLV5XUfg6PzztdQgItoCL5L98/wIuAPoBRwO7BsRO6eUXq1y2sER0Tul9Eqx309SkVJKvnz5amAvYAKQgEPycjPgrrxuOrBqXt8FuJvsP8ifA08CW+THbs3bF776AccCY4AvgXnAO8CpBfeuPG9QXu5aeX5Bm8rrdSVLRqreZ1A136lrwfFjgUlkieVvgabAd4CFwBdAy/ycdYCKwroq1xyQX2/UEv6OhfdcI69rD7yd1z2+pHZ5/ai8bkBeviAvzwG2LGjXFNikoLxj3u7T/HsmYPcqsfWs+nctOFb5Nz0rLw/Ky58BnQp+E8Py+n9X8z0qgIfz+rPyuqF1/dv25ashvBx2lBqBlNIC4JK8uCawU94L8wzQHxhN1gvTD3gmIjqQJWJj83OGk/WuTALWB94n6z25B+gMXB8ROyxjePeRJX8AT+X3+U8t51ySx9cKOJcs+fswr2tDNlQKcCAQwAMppTk1XG+9iLim4LXPkhqmlKYDV+fFPSKiVS2xFjowf38gpbSody2ltDAt3oN2bP4+BKgc/jtuKe5T1ffy94dSSp/m91wA/CWv37ma73EvcEBEbLMc95VUDZMvqfGYWPC5E7A/8F3gY7KenM+AD8iGsw5PKd0J/Ddv/3hK6ayU0jjgSrLerU/Ihus+JEtwdluWoFJK1wPj8uKd+X0er+W0w1JKPwIuyss/zN9vzt8rk5fKJOzOWq7XATiz4LV9Le0r/5bNyJLZYnWqcv63RERzsoQYssTr/vzzYUuZ6BXqkL9/UqV+cv7elGxItdCvyXo2L13Ge0paAud8SY3H+gWfPyUbtgJYjyzhKLRRDdd5GNi7mvqOS2jftKjolk5lj9xb+Xvl5PchZEOR+0REF7L5WVOAp2u53msppZ61tClU+bdcQJa0tik4tkrB5xb5+9f5+6dkf9vCfxZV7Us2tDmLbA7YArKh4vbAwWTDxEtrWn7ftarUr52/LwRmVDn+AVnP2Clkw7aSVhB7vqRGICKaARfnxc+AF8jmhQGMBJqklCKlFGQ9IL9ewnXW4JvEaxeyf4c8Vnk4f/8qf69MSLYoIsSF+Xux/07qlr9vlr9PAkgpzQf+BjQn6wVrCdyTUlr4rSsso4hYE/hpXnw6pTQ7pTSFLFkC2CFvtw6wQV5X2bP3SP5+aERsWXDNJhHx3bxY2Wu3OjAbmE+WeBUeW1pP5u8H50PKlb+JH+X1w1JKs6s573KyBwL6V3NM0jKy50tq2E6MiIPInnbchKwX5ccppa8j4p/AeGBr4IWIGE02Ab8fsB/ZpO2qviJLMlYnm8Q9k6x3qVDlU3PHR8QCFn9qb0k+zN/PjIgewF9TSq/V0P7+iHiWb5KC2wuO3QycDeyVl2sbcoR8zldB+aOU0pVV2vwmTz73JBvGmwqcUeW+ZwF3R8TLZIlhC7Kh28olOq4je8KwN/CfiHiUrFdsB+CxiPgl38wLG8o3PU6rk/2dvxcRHVNKU4v4ToWuJnsyciPg1Yh4iqzns1d+/3OqOymlNCkibgJWyPpukjL2fEkN2/7AEWRJwN+BnVK+zERK6Stgd7KnILsAxwObkk2kf7u6i+U9S8eTDUltT/aE5H1Vmt1OlvA0J1s+4Wpq9zuySf+bkw2BblxL+4vJeuDm5Of+oSDGt8h69gDeSykNL+L+Ved8VZcwDiT7e04Frgd6VZkkfyHZgwAfkC0lMRu4jeyJ04o8tq/IegwvI0s4DyablzaZrFfscLLeuo+BvVNKh6SUDsm/63tk/8N8ZBHfZzEppc/JErw/kD21eBzZP/P7ge1TSiNrOP1ysr+zpBUkUnKxaEkNS0ScD1wB/Co17PXJJK2ETL4kNRj5qvNHAaeSPUiwUUrpg7qNSpIW57CjpIZkI7JFV1cF/p+Jl6T6yJ4vSZKkMrLnS5IkqYxMviRJkspopVrnq0OHDqlr1651HYYkSVKtRo4cOS2l9K3dP1aq5Ktr166MGDGirsOQJEmqVURUu4+rw46SJEllZPIlSZJURiZfkiRJZbRSzfmSJEk1mz9/PpMmTWLOHLfkLJeWLVvSuXNnmjdvXlR7ky9JkhqQSZMm0bp1a7p27UpE1HU4DV5KienTpzNp0iQ22GCDos5x2FGSpAZkzpw5tG/f3sSrTCKC9u3bL1VPo8mXJEkNjIlXeS3t39vkS5IkrVARwdlnn72ofNVVVzFo0KBlvl7Xrl2ZNm1ajW3uvfdeunXrxm677bZM9+jXr9+itUQvv/zyZbpGsUo25ysingB6A9emlC6rcqwl8BegC/ABcGJKyZmBkiStYF0veHSFXm/CFfvX2qZFixbcf//9XHjhhXTo0GGF3n9J/vKXv3DTTTfRt2/f5b7W5Zdfzs9//vMVEFX1StnzdSJw7hKODQDeSintDLydlyVJUgPQrFkzBg4cyNVXX/2tYxMmTGD33XenR48e7LHHHnzwwQffajN9+nT23ntvunfvzkknnURKadGxO+64g2233ZaePXty8skns3DhQi699FKef/55TjzxRM4991wmTJjAzjvvTO/evenduzcvvvgiAEOHDuWAAw5YdK3TTjuNW2+9dbF7X3DBBcyePZuePXtyzDHHrKC/yOJKlnyllCbVcHhX4JH888N5WZIkNRA/+clPGDx4MF988cVi9aeffjrHH388o0eP5phjjuGMM8741rmXXHIJffv25c033+TQQw9dlKCNHTuWe+65hxdeeIFRo0bRtGlTBg8ezC9/+Uv69OnD4MGDufLKK+nUqRNPPfUUr7zyCvfcc0+191iSK664glatWjFq1CgGDx68fH+EJairpSbaAzPyz58Day6pYUQMBAYCdOnSpfSRDWpb+nuszAZ9UXubxs7fUM38DdXO31DN/A3V7uNX6/baqYI2s97jh4fuzXWXXUirVi3gq6/h41d56YXnuf/6/4GPX+W4PbbgvHPO/tY1n3v6Ce6/+Sr4+FX277Uu7dZoA5+M5umHnmDky/9hm55bADB7zlw6tVoIH28J82bB1Lfh46bMn/klp130v4wa8w5NmzThnfc/yO4x7V2YU/e/n7pKvj4D1sg/t83L1Uop3QjcCNCnT5+0pHaSJKl+Oeuko+m9z9GccMRBK+R6KcHxPziQ31x4eo3trr5pMGt1bM9rT91NRUUFLTfcAYBmzZpSUTCEWVcL0dbV047PAvvln/fLy5IkqQFZs11b+h+4F3+566FFdTv26cHdDz0BwOD7H2Pn7Xp967xdtu/NnQ88DsBjz7zAjM9nArBH322575F/8em0rM/msxlfMHHSx986/4uZs1inUweaNGnC7f94lIULFwKw/nrrMOad95k7dy6ff/45Tz/9dLVxN2/enPnz5y/HN69ZyZKviLiJbML9gIh4MCJ6RkTlBPxbgS0jYhiwZV6WJEkNzNknH8e0zz5fVP79Zefx13uG0GPP/tz+j0e59tJzvnXOxT8dyHPDX6H7bodz/2PP0GW9tQHYfJMNuey8U9n7qFPpsWd/9jrqFCZP+fYSFKce35/b7nuYrfY8grfGTWC1VVsB8J311qb/gXuxxRZb0L9/f3r1+nbiBzBw4EB69OhRsgn3UfgEQX3Xp0+fVLkGR8k416JmzrWonb+hmvkbqp2/oZr5G6rR2LFj6dbW1ZtqtG71SdfyGDt2LN26dVusLiJGppT6VG3rIquSJEllZPIlSZJURiZfkiRJZWTyJUmSVEYmX5IkSWVk8lVXmjSDE5+EX0zJntxZI1+9f78r4YKJWV3Po79pf9iN8POPsvdKJz4Fnbcpb9ySJGm5mHzVmQTvPAFvP7Z49SdvwPA/L163VnfY/BD428HQ/dCsvNVRMP1dmPRy+UKWJKkIq2+802LlW+8ZwmkXXbFC7zHod3/iqj/9rcY2U6fPYLsDfkivvY9i2PBXlvoet956K6eddhoADz74IGPGjFmmWKuqq+2FVLEQhv0Odv/F4vWv3AYb9lu8bsFcqJgP08bBwvnQpDns/DP4635IklSjG/ut2OsNHLpir1dCTz//X7bcbCNuvuqXy32tBx98kAMOOIDNN998ua9lz9fKYPo4ePH3cPwQePE66H4IjLwNdjkXjn8Ydvn26sCSJNVHDz/57KLeqD2P+DFTpk4H4NmXRtJzryPpudeR9Nr7KL6c9dW3zv31tTezSd9D6HvIj3j7vYmL6t+b8CH7HPMTtt7naHY+9Ee8NW48o954m/Muu5aHnhhKz72OZPbsOZxyweX02fcYunfvzsUXX7zo/K5duzJtWrZS/ogRI+jXr99i933xxRcZMmQI5557Lj179uS9995brr+BPV8ri6FXZK92G0D/2+CZy2Dr4+GPO8L5E+DNB2D68v0YJElaEWbPmUvPvY5cVP7s85kctPcuAPTdthf/efg2IoKb73yA395wG7+7+Gdc9ae/8YfLL2CnbXoy66uvadlilcWuOXL0GO4e8iSjnrqLBQsW0nufo9m6R7ai/MDzLuNPV1zExht2Yfgrr3Pqhb/hmXtv5NJzfsyI0WO4/tcXAPDr83/Cmu3asnCtHuyxxx6MHj2aHj161Pp9dtxxRw466CAOOOAADj/88OX++5h81aUOG0OrNbPP7TbIhhebNoc262V1q6+d1c8Y/8053/s1PPkLaL4qtGgDnbpD01WgaYvyxy9JUjVatWzBqKfuXlS+9Z4hjBidzZeaNHkKR5xyAZM/nca8efPZoMu6AOy0TU9+dsn/ccyh+3LYvrvTed21FrvmsOGvcug+u7Fqq2yfxoP22hWAWV99zYsjR/ODk89b1HbuvHnVxvX3h5/ixsH3syCaM3nyZMaMGVNU8rWimXzVpdMK9qk8fgiMGpw99dh156xuz4uhzwC4Jv9hfHf3bM7X+OeyJO2T1+GEf8Lr98KnK2YSoCRJpXT6//yWnw08loP23pWhL45g0P9lD5ldcNoJ7L9HX/75zAvsdMgJPHHnH9hsow1qvV5FRQVrtGm9WLJXnfEffMRVf/4bLz96B+0235UBAwYwZ062B2azZs2oqKgAWFRXSiZfdWlpN89975nsBVkSNnj5uz4lSSqnL2bOYr21OwJw270PL6p/b8KHbNltY7bstjEvj3qTt8ZNWCz52mX73gz46cVceNoJLFi4kIefeo6Tj/s+bVqvzgbfWZd7H36KHxy4FyklRo95l626b7LYfWd++RWrtWpF2zarM2XKFB577LFFc7u6du3KyJEj2XffffnHP/5RbdytW7fmyy+/XCF/AyfcS5Kkshl09sn84OTz2Xqfo+mw5hqL6q+5+U622P0H9NizP82bN2Pf3RZfrqL3lt044sC92WqvI9n32NPZpmf3RccGX/9r/nL3g2y15xF03+1wHnpy6Lfuu1X3Tei1xWZststhHH300ey00zfXv/jiiznzzDPp06cPTZs2rTbuI488kiuvvJJevXot94T7SCkt1wXKqU+fPmnEiBG1N1weS9sb1dgM+qKuI6j//A3VzN9Q7fwN1czfUI3Gjh1Lt7alHzpbqa3ba4VfcuzYsXTr1m2xuogYmVLqU7WtPV+SJEllZPIlSZJURiZfkiRJZWTyJUlSA7MyzeduCJb2723yJa2sOm4KPx4GF02G896HvX6V1Q94BC6YCBdOgqPuhpb500SH3Qg//yh7r3TiU9B5m/LHLqlkWrZsyfSvFpiAlUlKienTp9OyZcuiz3GdL2ll1awlvHY3jPsXbDsQdjoDxj0Fn46Fx38Oa28Bh/wRdjgVxjwEmx8Ct+4HJzwGL1wLa/eA6e/CpJfr+ptIWoE6d+7MpKcfY2rbDYGo63Dqpy/GrtDLtWzZks6dOxfd3uRLWllNfi17Abz/LGxzErRqB/88N6v74sPsvVW7bOuqivkwbVy2QG+T5rDzz+Cv+9VN7JJKpnnz5mzwnwvrOoz6rY6XKzH5klZ2LdrArudlG6u/++Q39XtfBgvmwMhbYfo4ePH32TZWL14H3Q+BkbfBLudCp24w/ll47qo6+wqS1JiYfFXRdc6ddR1CvTahrgPQ4lq0geMegFXbZ71Y82dn9QdcDT36w99/CFPezOqGXpG92m0A/W+DZy6DrY+HP+4I50+ANx/IEjhJUkk54V5aWbVoDT98ENp/F+7/f7BwblZ34LWw9QB47IJsWHLV9ouf971fw5O/gGiSJW+dukPTVaBpizr5GpLU2NjzJa2s1tkK1ts6+zzg0ex96G+yxAvggP8D/g8mDINbD8jqvrt7Nudr/HPQtDl88jqc8E94/V74dEy5v4EkNUomX9LKasLz1e8BOPSKJZ/z3jPZC7IkbPDhpYlNkrREDjtKkiSVkcmXJElSGZl8SZIklZHJlyRJUhmZfElSY9VyDTjtZbjoEzh/YrbvZ7MWsN+V2f6gg76Ankd/0979QaUVwuRLkhqrivnw9K+yhXZf+Rv0OAI2/h588gYM//Pibdfqnu0P+reDofuhWXmro9wfVFoGLjUhSY3VvK9g7JDs8xcfZttRTR+Xrfm2Yb/F27o/qKrTcg046Slo+53sN/LuEzDkdPj+X2DDXSGaZos9P3gqzBif9Zputj+89SjcPzC7xolPwRM/b1RJvMmXJDVmXXbItqhq3grGPQ2fT6y+nfuDqjqVvadT3oCtT4CdzoCxj8Doe+DpS7IdNPrfBtufAq/clvWe3rofnPAYvHAtrN2jUfaeOuwoSY3Zx6/Cn3eGZ34FG+0BvY5dctuhV8Cfd4HX7sl2S5j2Dnx3N7jjMOj702yrKzUulb2nn72/eO/p2Idh2rvZbwSyPWaX1Hv61MV1+x3qgD1fktRYrb1ltvfnjAkw7+usbv5saNsZ2qyXlVdfO9uMfcb4b86r3B+0+aruD6ol957+/CNYZfXs9zXxBXtPC9jzJUmN1Wod4aDfZ0889j0L/nsjjLoTDv0THHJD1mbPi+H4h745p3B/0Peedn9QLbn39E87w11HQet1YM9BWZ29p4A9X5LUeL33DFyz5bfrKzdiX9I57g+qStX1njZtAd0Pg8mjYN4sSBUw/+vFz2vkvacmX5Ikadms1hEOvBZarw2zZ2S9p6/dBcc/DGtukM0BG/8s/GvQN+cU9p42bd4oe09NviRJ0rJZUu/pDdvXfE4j7z11zpckSVIZmXxJkiSVUUmTr4gYEBEvRsQLEdG7yrENI+K5iBgaEf+OiM6ljEWSJKk+KFnyFRHtgDOAfsCxwHVVmpwK/CWl1A+4DTi9VLFIkiTVF6Xs+doWGJZSmpdSGg+0jojCZ0jfBNbIP7cDPi1hLJIkSfVCKZ92bA/MKCh/DqwJTM7L/wKeiIgTgRZkydq3RMRAYCBAly5dShasJElSOZSy5+szvunZAmib11X6X+AXKaUewCDg8uouklK6MaXUJ6XUp2PHjqWKVZIkqSxKmXwNB/pGRPOI6ALMSinNLTgewLT886dkvWKSJEkNWsmGHVNKMyLiBuBZIAFnRkRPYK+U0pXAZcCfI2IB0Bw4uVSxSJIk1RclXeE+pXQLcEuV6lH5sTeBvqW8vyRJUn3jIquSJEll5N6O0grWdc6ddR1CvTahrgOQpDpmz5ckSVIZmXxJkiSVkcmXJElSGZl8SZIklZHJlyRJUhmZfEmSJJWRyZckSVIZmXxJkiSVkcmXJElSGZl8SZIklZHJlyRJUhmZfEmSJJVRrclXRLSKiCsjYlRE7BQR10VE/3IEJ0mS1NAU0/N1DfBTYEugBdAUOLeUQUmSJDVUxSRfhwFXFpRHApuWJhxJkqSGrZjkqwKIgvJWwKzShCNJktSwNSuizaPAz/LPtwNrAzeXLCJJkqQGrJjk6yyynq/9gebAbTjnS5IkaZnUmnyllGYCJ5QhFkmSpAav1uQrIp6ppvpz4KmU0h9XfEiSJEkNVzHDjv2WUH9wRHRIKf1qBcYjSZLUoBXztOOvgYeBTciWmHgYuBq4Ezi+dKFJkiQ1PMX0fP0E+E1KaRxARAwDzgaOBQ4vYWyS1Ch1nXNnXYdQr02o6wCk5VRM8vUR8OuIOBBIwA7AWKA9ML2EsUmSJDU4xSRfR5MtL9E3L78KDADWBM4oTViSJEkNUzFLTbwO9I6INnl5ZsmjkiRJaqCKWWoigCPJNtZumRVJKaWzSxybJElSg1PMsOMfgB+Tzfeq3OMxkU26lyRJ9YwPbdRsQh3fv5ilJg4lW1YC4Ezg34Bre0mSJC2DYpKvdsCw/PNk4D5gYMkikiRJasCKGXb8JG/3CXAzsArgpHtJkqRlUEzP1y+A94CfAXOAL8iGHyVJkrSUiun5agKMTSlNBO6JiPZk2wxJkiRpKRXT8/VXYNuC8l58MwdMkiRJS2GJPV8RcRBwCNnyEqdGxL75oV5kw4+SJElaSjUNO/Yi20YoAbvmr0qDSxiTJElSg1VT8nUj8CjwX+Ai4EmyRGxGSml8GWKTJElqcJaYfKWUJpOt69UkIpoBawFNASKiS0rpg/KEKEmS1HAUs7fj6cAVQMuC6lTMuZIkSVpcMQnUJWQT7J8DFpQ2HEmSpIatmOTrA+DPKaU/ljoYSZKkhq6Y5GsM8D8RsS4wI69LKaWrSxeWJElSw1RM8nVk/n5RQV0Cak2+ImIA2SbcCTg9pfRKlePnky3a2gy4NKX0TBHxSJIkrbSKSb5OWJYLR0Q74Axge2A94Hagb8HxfYG2KaU9l+X6kiRJK6Nak6+U0m0RsQqwOTA+pfRFkdfeFhiWUpoHjI+I1hHRIqU0Nz/eH5gREU8DHwOnLcW1JUmSVkq17u0YEb2A94ARwDYRMTYibiri2u35Zo4YwOfAmgXldYGKlNIewHDgwiXcf2BEjIiIEVOnTi3itpIkSfVXMRtrXw98RbbHYwVwB1DMUOFnwBoF5bZ5XeHxx/PPjwM9qrtISunGlFKflFKfjh07FnFbSZKk+quY5Gsr4NaC8sdApyLOGw70jYjmEdEFmFUw5AgwFOiTf+4DjCvimpIkSSu1YibcT+KbTbV7AEcBE2o7KaU0IyJuAJ4le9rxzIjoCeyVUrqSLKG7KSL+DcwHfrjU0UuSJK1kikm+fgvcnH/+Hdnw44BiLp5SugW4pUr1qPzYXEy4JElSI1PM0463RMR7wP551aMppWdLG5YkSVLDVMzG2tsB81NK5+XlHSNiu5TS8JJHJ0mS1MAUM+H+Xr6Z8wWwS14nSZKkpVRM8tWebI2uSjNZfL0uSZIkFamYCfdvAb+IiEQ22f4iYGxJo5IkSWqgikm+LgIeAP5AlnzNBX5UyqAkSZIaqmKSrxeBbcjmegE8mVJyQVRJkqRlUGPyFRFBtqL9z1JKN5QnJEmSpIarxgn3KaUE/JNvtgGSJEnScihm2LEn8N2IOBiYktellNJWpQtLkiSpYSom+doof++YvyRJkrSMitleqJi1wCRJklSEohKriNg/Iq6JiM0j4tiIcMhRkiRpGdSafEXEWcDDwOnA2sBhwJUljkuSJKlBKqbn6ywW38vxX0Dv0oQjSZLUsBWTfLUDXisorwo0LU04kiRJDVsxTzsOB07JP58D9AVeKFlEkiRJDVgxPV9nALPJ9nXcB5hMNhQpSZKkpVTMUhNvRUQ3YNO86u2U0sLShiVJktQw1Zp8RURb4H+AfnnVvyPispTSF6UMTJIkqSEqZs7XncC+QEVe7g10Aw4oVVCSJEkNVTFzvnYEbgRakT3peHNeJ0mSpKVUTM/XI8DUlNJ8gIiYCgwpaVSSJEkNVDHJ17bA0RHxw7zcGXgnIkYDKaXkVkOSJElFKib52jh//05B3abVNZQkSVLNillqoqjNtyVJklQ7EytJkqQyMvmSJEkqI5MvSZKkMqo1+YqIH0bE+gXl9hHhOl+SJEnLoJier7+SLTdRaS9gWGnCkSRJatiW+LRjRBwEHAIEcGpE7Jsf6gXMKUNsDdoqTZvwm+9vyZ7d1qJJwH/en85P73mNXl3W4OIDu9OpTQuef3caF9w/mpmzF3D+Ppty3A5deWXiDE66bQTzFlZw/dG9eOKNT3h49OS6/jqSJKlINfV89QIGAAnYNf88ANgKuL/EcTV4u27ake/37sztL03k8n+OZa/N1+YHfTpz/dG9eX/qLI65aTh9N+7A+d/bjNYtmnFKv4348e0j2XTt1uyySQe222BN2q/WwsRLkqSVTE3J141kw40BXARsA/QBvptSOq4MsTVoH0z/mrkLFvLJF7P5aMZsIOsNa9uqOc+9M5XXP/qCd6fMYrfNOrGgIjF/YQXvT53F7HkLWbAwcdH+3Rg05M06/haSJGlp1bTI6lzgfaBD1QMRsWZK6bOSRdUITJj+Fc+9M5XLDt2ShRWJ4eOn88joyZy/z2Z0W6cNq63SlO+0a0Xrls2ZPX8hg4a8yZ+P68Pz44JG5scAABKPSURBVKbRec1VGTlxBnt1X4tBB3Xn7SlfmohJkrSSqCn5mkY25FidVMu5qsXhW3dmr83X5rePv8WHn33N74/uzX5brsM1T7/DmXtswhHbfIev5i3k48+zXrHBwz9g8PAPaNOyGfedsiPn3TeaO07ajp6XPMmTP92FnTfuwLB3p9Xxt5IkSbWpKYF6jiUnX1pOFSn7086Zv5C5CyoAWLttCx589WOefHMK67dfjUEHbc69Iz9c7Lyf7rUJNw17n9nzFrJK0yZstk5rVm/RjFWauWSbJEkrgyUmXymlfmWMo9G5/5WP2PG7HThrz01o2iR4Ydw0bnpuPBcftDm7b9qJz76exwOvfMTNw8YvOmejTquz5XptueThMQAMee0j/n7yDvx3/Gc8987UuvoqkiRpKdQ6dBgRv6ymOqWUflWCeBqNuQsqOP2uV79Vf8odryzxnHGfzuLwP720qHzOvaM5597RJYlPkiSVRjHztgZVU5cAky9JkqSlVEzy1Z9v5n61A34CvFyyiCRJkhqwWpOvlNJ9heWI+Ap7vSRJkpZJMXO+CicVNQM2AGaWLCJJkqQGrJhhxy0KPlcAE4DzSxKNJElSA1fr4lAppSYFr2YppY1SSv8o5uIRMSAiXoyIFyKi9xLaXBIR45Y2cEmSpJVRUavUR8QOwJZAy8q6lNJ1tZzTDjgD2B5YD7gd6FulzVrAJksXsiRJ0sqr2HW+Lq7mUI3JF9mm3MNSSvOA8RHROiJapJTmFrT5H+A3wP3FBiwpc/jWnbnqB1stVvf3ER/y0nvT+emem7BWmxZM/OxrLn7oTV56fzrn77Mpx+3QlVcmzuCk20Ywb2EF1x/diyfe+ISHR0+uo28hSY1PMXvS/D/g8fzzb4DXgauKOK89MKOg/DmwZmUhIjYGVk8p1bhKaEQMjIgRETFi6lRXcZcqPTL6Y7a//Gm2v/xpLs43Vn/+3WkMOqg7cxYs5MDrn2fN1VbhFwd0o3WLZpzSbyN+fPtINl27Nbts0oHtNliT9qu1MPGSpDIrJvnqBDySfx4O3AAcXsR5nwFrFJTb5nWVBlHEkhUppRtTSn1SSn06duxYxG2lxmHO/Ao+mTmHT2bOYd8t1mbql3N47I3JjPt0FrPmLGDCtK+ZM38hs+YuYEFFYv7CCt6fOovZ8xayYGHiov27MShP2iRJ5VNM8jWN7CnHz4HfA5eQLbZam+FA34hoHhFdgFlVhhw3BP4QEY8D60REbcOYkqqxcafV2X7D9tz53w+ZvzAxePhEtuzcljGXfo+2rZpz2SNjmT1/IYOGvMmfj+vD8+Om0XnNVRk5cQZ7dV+Lu/7f9gw6qHtdfw1JajSKSb6uAb4gS7q+A3QALq3tpJTSDLJesmeBu4CzIqJnRJybH98hpbRPSmkfYHJK6Yxl/A5So/bDHboyf2EFdw6fyOotmnHFYT145YMZ9P/zS0z+Ys6ieWGDh3/Agdc/z28ff4sf7rA+D436mB/v+l2O+8twdtm4Aztv3KGOv4kkNQ5LnHAfEZcC9wBTgJdSShMj4mYgUkpfFXPxlNItwC1VqkdV026j4kOWVGm1VZpySK91eWrMFKbMnEvL5k2IgAULE3PmV7CwIrHuGq0WO+ene23CTcPeZ/a8hazStAmbrdOa1Vs0Y5Vmxfy/mCRpedX0b9uLgF7AX8meXCSl9HWxiZek0jusd2dat2zO7S9NBLJ5YL948A3Wb78qD/xkR1o2b8rP7399UfuNOq3Oluu15d4Rk3h7ypcMee0j/n7yDoyZPJPn3vGBFkkqh5qWmpgO/AEI4PqI+N+CYyml9N2SRiapVrf/ZyK3/2fiYnX3vPwh97z8YbXtx306i8P/9NKi8jn3juace2t84FiStILV1PN1OVA5Qb4N0LHg1anEcUmSJDVIS0y+UkrXpJQ6kU2Y3y+l1LrwVb4QJUmSGo5i9nbcDZgeEWdFxIYRsXNEfKcMsUmSJDU4xWwvdAQwmGzu12jgQuAr4JDShiZJKrXN12nDJQd3p/u6bZg5ewGn3fkKrVZpysUHdqdTmxY8/+40Lrh/NDNnL3CLKmkFKebZ8kuBpwvKjwA7lCYcSVK5rLZKU2770bYAHHbDi/ziwddp0iS4/ujevD91FsfcNJy+G3fg/O9t5hZV0gpUTPK1LvBMQXkB0GoJbSVJK4ndN+tEx9Yt+O3jb/HWJ1/yr7GfMnveQtq2as5z70zl9Y++4N0ps9hts05uUSWtQLUOO5JtpP3D/PNxwD7AayWLSJJUFuu1y/4/+vx9NqNrh9UYO3km1/7rXSoqEt3WacNqqzTlO+1a0bpl8xq3qBp0UHfenvKliVgjVd3Q9X2n7LhYmw+mf80uV/7boetcMcnX2WRDjQEcT7Y59jmlDEqSVHozvp4PwKsffs5vn3ibO0/ajs92mMc1T7/DmXtswhHbfIev5i3k489nA9kWVYOHf0Cbls2475QdOe++0dxx0nb0vORJnvzpLuy8cQeGvTutLr+Syqxy6HrC9K847IYX6dyuFXMWLGT7y7PZSmu1acEDp+7E8+OmLhq6Pvbm4fyu/1bsskkHvpyzoFEOXdeafKWUXoqIjfhmntdL+b6NkqSV2PPvTmPeggrmL6hg7vyFJGDuggr+/dZUnnxzCuu3X41BB23OvSMXX7TXLapUqXLo+tTBI3nrky9565MvFzt+5LbfoUmT4G8vTVzi0PW5jXCh5xqTr4joAJwG9AESMBJ4uQxxSZJK7KPPZ3POva9x9t6bcPyOXXn+3Wlc+cTbXHpwd3bftBOffT2PB175iJuHjV90TuUWVZc8PAZg0RZV/x3/mVtUNULVDV1feP/rTJoxm2ZNgqO27cLw8dMXJWUOXWdq2li7K/ACsDbZkCPAfsBJEbFjSumDkkcnSSqpIa99zJDXPl6s7pQ7Xllie7eoUqHqhq7P/d6mnHn3KL7XfW3WatOSXz0yZlF7h64zNfURXwasA9wMHAZ8P/+8DvCr0ocmSZLqsyUNXQMct8P6fDpzDo+/8cm3zmvsQ9c1DTvuAtyWUjq5oO6BiFgF2L20YUmSpPpuSUPXG3dane03bM+1/3qHBRVpsXMcuq45+eoI/Lua+qHAkSWJRpIkrVSqG7qe+uVcul7waLXtHbquOflqAZwREYdXqV8XWKV0IUmSJDVctS010Tt/VZWqqZMkSVItakq+NihbFJIkSY3EEpOvlNLEcgYiSZLUGDSOZzolSZLqCZMvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyKmnyFREDIuLFiHghInpXOXZeRAzPj/0+IqKUsUiSJNUHJUu+IqIdcAbQDzgWuK5KkwdSStullHYC1gJ2L1UskiRJ9UUpe762BYallOallMYDrSOiReXBlNK7BW3nAgtKGIskSVK9UMrkqz0wo6D8ObBm1UYRsSuwDvBcdReJiIERMSIiRkydOrUkgUqSJJVLKZOvz4A1Cspt87pFIqIHcAVwZEopVXeRlNKNKaU+KaU+HTt2LFmwkiRJ5VDK5Gs40DcimkdEF2BWSmlu5cGI2Ai4hSzxmlbCOCRJkuqNkiVfKaUZwA3As8BdwFkR0TMizs2bXEPWM3ZbRAyNiP1LFYskSVJ90ayUF08p3ULWu1VoVH7sgFLeW5IkqT5ykVVJkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyMvmSJEkqI5MvSZKkMjL5kiRJKiOTL0mSpDIy+ZIkSSojky9JkqQyKmnyFREDIuLFiHghInpXOdYyIgZHxLD8vWUpY5EkSaoPSpZ8RUQ74AygH3AscF2VJgOAt1JKOwNv52VJkqQGrZQ9X9sCw1JK81JK44HWEdGi4PiuwCP554fzsiRJUoPWrITXbg/MKCh/DqwJTK7meOWxb4mIgcDAvDgrIt5e8aHWax2AaXUdRKX437qOQMvA35CWl78hLa/G+htav7rKUiZfnwFrFJTb5nXVHa96bJGU0o3AjaUIcGUQESNSSn3qOg6tvPwNaXn5G9Ly8je0uFIOOw4H+kZE84joAsxKKc0tOP4ssF/+eb+8LEmS1KCVLPlKKc0AbiBLqu4CzoqInhFxbt7kVmDLiBgGbJmXJUmSGrRSDjuSUroFuKVK9aj82GzgqFLev4FotEOuWmH8DWl5+RvS8vI3VCBSSnUdgyRJUqPhCveSJEllZPIlSZJURiZfkiRJZVTSCfdaehGxGXAwsF5e9REwJKU0tu6iktSY5P8eWg8YnlKaVVC/T0rp8bqLTCuLiNgWSCmllyNic2Afsi0F/1nHodUL9nzVIxFxPnA3EMB/81cAd0XEBXUZm1Z+EXFCXceg+i8izgAeAk4H3oiIgwsOX143UWllEhEXk+3n/MeI+A1wPbAacEFEXFSnwdUTPu1Yj0TEO0D3lNL8KvWrAG+mlDaum8jUEETEBymlLnUdh+q3iHgd2CGlNCsiugL3AbenlK6NiFdTSr3qNEDVe/lvqCfQAvgE6JxSmhkRrch6U3vUaYD1gMOO9UsFsC4wsUr9OvkxqUYRMXpJh4C1yhmLVlpNKocaU0oTIqIfcF9ErE/2O5JqsyCltBD4OiLeSynNhGx9z4jwv2WYfNU3ZwFPR8S7wId5XRdgI+C0OotKK5O1gO+x+Kb2kP1H88Xyh6OV0JSI6JlSqlwQe1ZEHEC2YPaWdRuaVhLzImLVlNLXwNaVlRHRFjsSAIcd652IaAJsy+IT7l/O/y9CqlFE/AX4a0rp+WqO3ZlSOroOwtJKJCI6k/VcfFLNsZ1SSi/UQVhaiUREiyp7OVfWdwDWSSm9Xgdh1SsmX5IkSWXk046SJEllZPIlSZJURiZfkupcRHSNiFTl9XlEDCgofxURb0fElRGxapXzHim41iN5Xde8vHpEXBMRkyJiTkS8ExE/LmjfLiLm5ef8Jq8bVE08la9+ETE0/9whb98+Iv4aEVMjYm5EjIqIHxTc49a8/dCCugkRsWgBU0mNh8mXpPrkVeCo/PWjgvrBwNnAFOAc4OGIqHXZg7zNI8CZwBiyhUPvBbYpaHYo0JzsKazKhOm+PIaBeXlsQVxjqrnV/cAAYAjwM6A98PeI2L1Ku10jYofa4pbUsLnUhKT6ZCrwr/zzfLLECGBUSulPEXET8AawO7Ab8H4t19sd2JUsYdonpVQBi54qrtQfmAvcDPwkIvqklEaQre7eAbgR+DSldHflCYV5X74O1i7AiymlE/O6ccDjwC+BZwruNRO4CDig1r+EpAbLni9J9cneZAnYVLItbhaTL7nyZF7sU8T1KtcYeqoy8cqvU5mEtQf2yK/51/xw/6WMuXLtq5cK6irXVNuiSts/AftHRKNf4VtqzEy+JNUnw4G98tfZS2hT2e2UqH7BxsrjFXmbmhxGNgLwAjCdbCuUpU2+lsY/gLeBn5fwHpLqOZMvSfXJtJTSv/LXyKoHI6IpsGdeHEk2ByyRbcFVaV2yxOuTvA3AnoVDjQWfj8jfrwDGA2sD60fEdksRc+WCkdsX1FXO63qjStsK4H+Bw4EOS3EPSQ2Ic74k1SfrRsSRBeXm+XvPiDiZbMJ7N7J5VP9OKaWI+Dewe0TcQtbr1ZNsmHFefmwo0A/4Z0TcR7Zl17oR8fO8fjhZ8gWwAfB/ZEnZ8GICTikNjYjngF3yOWmvAufnhy+t5pQ7gEF5HF8Vcw9JDYs9X5Lqk17AXQWv1fL6Y4CryXqmfgscmL7ZnuMEsvlhBwMHAQ+SPymZtzkQuI5sbtYN+bVGAt8HmgL3ppQeTCk9CPyebFL8D4p5mrLAYcCtZA8IXA18DvRPKT1TtWFKaT5w1VJcW1ID4/ZCkiRJZWTPlyRJUhmZfEmSJJWRyZckSVIZmXxJkiSVkcmXJElSGZl8SZIklZHJlyRJUhmZfEmSJJXR/wfsdDXdvEuiVgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Plot a stacked bar chart to show default by education level\n",
        "\n",
        "create_stacked_bar('EDUCATION')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZguFpYvFAmz6"
      },
      "source": [
        "#### (3) Did customers with a high education level get higher credit limits?\n",
        "From the boxplot, we can see that customers with grad school education have the highest 25% percentile,highest median,highest 75th percentile and highest maximum numbers, which proves that customers with higher education levels did get higher credit limits."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 331
        },
        "id": "JoxIqbUDAmz6",
        "outputId": "2d7fb0ce-c73b-4e54-8747-4c0b0e386488"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEDCAYAAAAlRP8qAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3df3xV9Z3n8dcnMT+UHxJHBm1QQxGnIWl1SratkbFmxlqjM2h3bRW1awvVHbukU7vt0ja79cc84FG2wHZEpxYqY6eFgLWjw4oM1U2sA6gVWlCSDDUKVmG1VKP8miT8+Owf9yTeCzc/7w3nnpP38/G4j5wf33vv535z8sn3fs/3fI+5OyIiEn15YQcgIiLZoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE6EmdDNbbma/N7PtAyz/OTNrMbNmM1s53PGJiESJhTkO3cwuBQ4A/+julf2UnQI8DPy5u7eb2R+7++9PRpwiIlEQagvd3Z8B3kneZmaTzexfzGyLmf2rmX0o2HUrcL+7twfPVTIXEUmSi33oS4E6d58GfB34+2D7BcAFZrbRzJ4zsytDi1BEJAedEnYAycxsNFAN/MzMujcXBT9PAaYAlwETgWfM7MPu/u7JjlNEJBflVEIn8Y3hXXe/KM2+N4Dn3f0wsNPMfksiwb9wMgMUEclVOdXl4u77SCTrzwJYwoXB7sdItM4xszNJdMG8GkacIiK5KOxhiw3As8CfmNkbZjYbuAmYbWbbgGbgmqD4euBtM2sBmoBvuPvbYcQtIpKLQh22KCIi2ZNTXS4iIjJ0oZ0UPfPMM72srCystxcRiaQtW7b8wd3Hp9sXWkIvKytj8+bNYb29iEgkmdlrve1Tl4uISEwooYuIxIQSuohITCihi4jEhBK6iEhM9JvQ+7sJRXB5/r1m1mZmL5rZR7MfpkRdQ0MDlZWV5OfnU1lZSUNDQ9ghicTOQFroDwF9TVVbS2KSrCnAbcAPMg9L4qShoYH6+nqWLFlCR0cHS5Ysob6+XkldJMv6TejpbkJxnGtI3HHI3f05YJyZnZ2tACX65s2bx4MPPkhNTQ0FBQXU1NTw4IMPMm/evLBDE4mVbPShlwKvJ62/EWw7gZndZmabzWzz3r17s/DWEgWtra1Mnz49Zdv06dNpbW0NKSKReDqpJ0Xdfam7V7l71fjxaa9clRgqLy9nw4YNKds2bNhAeXl5SBGJxFM2Evpu4Jyk9YnBNhEA6uvrmT17Nk1NTRw+fJimpiZmz55NfX192KGJxEo25nJZA8wxs1XAx4H33P3/ZeF1JSZmzpwJQF1dHa2trZSXlzNv3rye7SKSHf3Ohx7chOIy4EzgLeBOoADA3R+wxM0/7yMxEuYQ8EV373fWraqqKtfkXCIig2NmW9y9Kt2+flvo7t5nM8oT/xH+6xBjExGRLNGVoiIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMTGghG5mV5rZDjNrM7Nvptl/rpk1mdlvzOxFM7sq+6FKlDU0NFBZWUl+fj6VlZU0NDSEHVJkqS6lV+7e5wPIB14BPggUAtuAqceVWQrcHixPBXb197rTpk1zGRlWrlzpkyZN8sbGRu/q6vLGxkafNGmSr1y5MuzQIkd1KcBm7y1f97bD30/WFwPrk9a/BXzruDI/BOYmld/U3+sqoY8cFRUV3tjYmLKtsbHRKyoqQoooulSX0ldCt8T+3pnZdcCV7v6lYP3zwMfdfU5SmbOBXwAlwCjgcnffkua1bgNuAzj33HOnvfbaa4P6NiHRlJ+fT0dHBwUFBT3bDh8+THFxMUePHg0xsuhRXYqZbXH3qnT7snVSdCbwkLtPBK4CfmJmJ7y2uy919yp3rxo/fnyW3lpyXXl5OXfffXdKv+/dd99NeXl52KFFjupS+jKQhL4bOCdpfWKwLdls4GEAd38WKAbOzEaAEn01NTUsWLCAWbNmsX//fmbNmsWCBQuoqakJO7TIUV1KXwaS0F8AppjZJDMrBG4A1hxX5nfAXwCYWTmJhL43m4FKdDU1NTF37lyWL1/OmDFjWL58OXPnzqWpqSns0CJHdSl96bcPHSAYhvh9EiNelrv7PDO7h0Tn/BozmwosA0YDDvx3d/9FX69ZVVXlmzdvzvgDSO5Tv2/2qC4l4z50d3/C3S9w98nuPi/Y9h13XxMst7j7Je5+obtf1F8yl5GlvLycDRs2pGzbsGGD+n2HQHWZfbEa19/b8JfhfmjY4sihsdPZo7rMrijWJ5mMQx+uhxL6yHLFFVe4mTngZuZXXHFF2CFF1pw5c7yoqMgBLyoq8jlz5oQdUmRVVFR4fX29V1RUeF5eXsp6ruoroWsuFxl2dXV1NDY2snDhQg4ePMjChQtpbGykrq4u7NAip6GhgbVr17Ju3Tq6urpYt24da9eujXY3QYhaWlpYsWIFS5YsoaOjgyVLlrBixQpaWlrCDm1oesv0w/1QC33kKCoq8ptuuimlFXTTTTd5UVFR2KFFThRblLmsqKjIFy1alLJt0aJFOX1soi4XCRPgZWVlKf2UZWVlnmhPyGCYWdo+XzMLO7RIimJ99pXQ1eUiw87MqK2tpaamhoKCAmpqaqitrcXMwg4tcgoLC6murqauro7i4mLq6uqorq6msLAw7NAiaerUqVx44YXU1tZSWFhIbW0tF154IVOnTg07tCFRQpeTYunSpSxevJhDhw6xePFili5dGnZIkdTZ2cnq1atTrhRdvXo1nZ2dYYcWSTU1NTz++OPMnz+fgwcPMn/+fB5//PHIXnk7oAuLhoMuLBo5KisrOXToEDt37uzZNmnSJE477TS2b98eYmTRU1xcTFVVFZs3b6azs5OioqKe9Y6OjrDDi5zKykquvfZaHnvsMVpbWykvL+9Zz9Vj82RMziXSq9LSUnbu3Mntt9/Ou+++y+23387OnTspLS0NO7TI6ezs5Pnnn09pUT7//PNqoQ9Ra2srd955J9u3b+fo0aNs376dO++8k9bW1rBDG5reOteH+5HrJ0VXrlyZMpIgly80yHVFRUV+ySWXpIyd7l6XwSkqKvLTTz/dSUyx4YCffvrpqsshiuL88miUy+BE8eqxXAb4eeedl1Kf5513nka5DEF3Eq+urvY9e/Z4dXV1zzYZvCj+rSuhD1IU/2vnMjPzkpKSlFZlSUlJTg8Ny1WAFxQUpNRl97oMTdS+jfeV0HVSNA3NaJdd3cMTKyoqeOKJJ7jqqqtobm4GEg0KGbjuuqyuruaRRx7huuuuY9OmTYDqcqTQSdFB0ox22XfWWWfR1tbGeeedR1tbG2eddVbYIUXW2LFjee+995g4cSLvvfceY8eODTukSIvTbItK6GnU19cze/ZsmpqaOHz4ME1NTcyePZv6+vqwQ4uswsLClPlHdCHM0O3bt4/Jkyfz1ltvMXnyZPbt2xd2SJHV0NBAfX19ylwu9fX10U3qvfXFDPcjl/vQ3aPXr5bLAB8zZkxKv2/3ugwO4MXFxSl12b0ugxfF82WoD13CNHr0aA4ePEhZWRlPPfUUl19+Obt27WLUqFEcOHAg7PAi5dxzz+X1118/oQ/9nHPO4Xe/+13Y4UVOFM+X9dWHfsrJDkZGnkOHDnHKKaewa9cuzj//fABOOeUUDh06FHJk0TN27FgmTZrEpk2b+MAHPgC8f9WtDF73+bLkS/2jfL5Mfegy7NydI0eOpFwpeuTIEY3KGILW1lZ27NiR8jV7x44d0b2yMWRxO1+mFrqcFCUlJSxfvpwf/OAHFBUVUVJSQnt7e9hhRU55eTl33333CXOPRLVFGbaZM2cCiZuwdNfnvHnzerZHjVroclK0t7czbdo09uzZw7Rp05TMh6impoYFCxakzLa4YMGCyM4OmAtmzpyZMpdLVJM5KKHLSVJQUNDT77tp06aUk1AycE1NTcydO5fly5czZswYli9fzty5c2lqago7tMiK0zh0DVuUYYfmH8mavLw87+rqStnW1dXleXl5IUUUbXGby0XDFmXY9XVnorCOv6iK4vzduSyK9alL/yUnlJWV0dbWRllZWdihRJb60LOrpaWFFStWpFwpumLFClpaWsIObUjUQpdhZ2aYWUprvHtdLfTB0d2fsqu4uJj58+fzta99rWfb4sWL+fa3v52zd4BSC11C5+5UVFTw2muvUVFRoUQ+RM3NzWnv/tQ9e6UMTldXF/fdd1/KOPT77ruPrq6usEMbErXQZdipDz17zIwxY8awf//+nm3d66rLwVMfusgQjR49mi1btjB69OiwQ4m0/fv3U11dzZ49e6iurk5J7jI49fX1rFy5MqUPfeXKlZG9UlQtdBl2aqFnj5lRVFREV1dXYpiaGYWFhXR2dqouh+jTn/40Tz75ZE99fupTn2L9+vVhh9UrtdAlJxQXF/Pcc89RXFwcdiiR1tnZycUXX8yePXu4+OKL6ezsDDukyKqrq6OxsZGFCxdy8OBBFi5cSGNjI3V1dWGHNiQDaqGb2ZXA3wH5wI/c/btpynwOuIvEBSPb3P3Gvl5TLfSRQy307FFdZldxcTElJSW8+eabPdvOOuss2tvb4znKxczygfuBWmAqMNPMph5XZgrwLeASd68Avppx1BI7yf2+kpmSkhJefPFFSkpKwg4l0jo7O3nzzTeZMWMGe/fuZcaMGbz55puR/dYzkNkWPwa0ufurAGa2CrgGSB55fytwv7u3A7j777MdqERbXl5eyhzeeXl5HDt2LOSooikvL4/29nY+8pGP9KyrLoeutLSUV155hQkTJlBeXk5paSm7d+8OO6whGUgfeinwetL6G8G2ZBcAF5jZRjN7LuiiOYGZ3WZmm81s8969e4cWseSk7ouH0j2AExJO93pfz5H0jh07ljIOXck8M7t37+bSSy/lnXfe4dJLL41sMocB9KGb2XXAle7+pWD988DH3X1OUpnHgcPA54CJwDPAh9393d5eV33oI0v3bei66fZzQ6OrbrPLzBg7dmzPOP7ucf779u3L2frMdJTLbuCcpPWJwbZkbwBr3P2wu+8EfgtMGUqwEk8HDhzo+QNxdyXzDLh7Sh96riaeqNi3b1/KqKF9+/aFHdKQDaQP/QVgiplNIpHIbwCOH8HyGDAT+AczO5NEF8yr2QxURBLy8/NT+tDz8/Nz9obGua777lnJ53e6R7lEUb8tdHc/AswB1gOtwMPu3mxm95jZjKDYeuBtM2sBmoBvuPvbwxW0yEi3aNEiDh48yKJFi8IOJdK6uro49dRTaWxspKuri8bGRk499VTN5TJY6kMfmY7v/5XBUR96dmkuFxEJlbunzIujRD50cZvLRQm9F3V1dRQXF2NmFBcXR/ZSYImnAwcOMG3aNJ1cztDMmTO5+uqrqa2tpbCwkNraWq6++urI3ihaCT2Nuro6HnjgAebPn8/BgweZP38+DzzwgJK65IwJEybQ2trKhAkTwg4l0hoaGli7di3r1q2jq6uLdevWsXbt2sjeKFp96GkUFxdz3XXXsXXr1p5+tYsuuohHHnkkZ+d3iAr1oWem+6KroqIiOjs7e36C5nIZCvWhjwCdnZ1s3LgxpV9t48aNkZ3fQeLn7LPPpq2tjbPPPjvsUCKtpaUlbR96VO8pqoSehplRW1tLTU0NBQUF1NTUUFtbq0vSJWfs2rWL888/n127doUdSqQVFhZSXV3dc86srq6O6upqCgsLww5tSJTQ03B3li1bxuLFizl06BCLFy9m2bJl+korOUN3f8qOrq4uVq1axaxZs9i/fz+zZs1i1apVGoc+WLnch15ZWcmUKVNYt25dTz9lbW0tL7/8cs72q0WF+tAzo/nQs2vEzYc+EtXX17Nt27aUM9/btm2L7NhUiR/NLZ8dcZsPXS30XtTV1bFs2bKeFvqtt97KkiVLwg4r8tRCz4zuKZpd3deZJLfGu9dztT7VQh+kuI1NlXjRPUWzq6OjI+UbT652tQyEWuhpVFZW8sorr5zwX3vy5MnqQ8+QWuiZUR96dkWxPtVCH6Tm5mY6OjooKyujra2NsrIyOjo6aG5uDjs0ERkmF198cdghZEwJvRdjx45l1KhRXHDBBYwaNYqxY8eGHZKIDKNnn3027BAypoTei3379qWMTY3yXUwknu69996wQ4iNuMyNoz70NLrPfLt7zygXM8vpM99RoT70zESxzzeXRbE+1Yc+BB0dHYwbN47W1lbGjRsX6TPfItK/iy66KOwQMjaQe4qOOEVFRRw7doy33nqL8vJyAAoKCsjL0/8/kbjaunVr2CFkTBkqja6uLkpLS1PuM1haWhrZ+R0knj75yU+GHUJsFBcX89xzz1FcXBx2KBlRCz2NqVOncu2111JXV9czR/JNN93EY489FnZoIj1++ctfhh1CbHR0dPCJT3wi7DAyphZ6GnG7z6DE02c+85mwQ4iN/Px8nn76afLz88MOJSNqoafRfT/B5Bb6vHnzInufQYmnRx99NOwQYuPo0aNcdtllYYeRsRHdQjezXh833ngjzc3NHDt2jObmZm688cY+y4ucbHfccUfYIcTKT3/607BDyJjGofdD46azS/WZmSiOm85lUaxPjUMXiZihfBPUN8ih0ygXERk2fbUO0yXoXG1NRoVGuYhIKNy9J4EnL0tmVq1aFXYIGVNCFxEBbrjhhrBDyJgSuogI8NBDD4UdQsaU0EVEgC984Qthh5AxJXQRGfEKCgrYsGEDBQUFYYeSkQEldDO70sx2mFmbmX2zj3L/yczczNKOkRQRyUWHDx9m+vTpHD58OOxQMtJvQjezfOB+oBaYCsw0s6lpyo0B/gZ4PttBiogMt1tuuSXsEDI2kBb6x4A2d3/V3buAVcA1acr9LbAA0J0gRCRyfvzjH4cdQsYGktBLgdeT1t8ItvUws48C57j72r5eyMxuM7PNZrZ57969gw5WRGS4xGEul4xPippZHrAY+G/9lXX3pe5e5e5V48ePz/StRUSy5uabbw47hIwNJKHvBs5JWp8YbOs2BqgEnjazXcAngDU6MSoiUXL99deHHULGBpLQXwCmmNkkMysEbgDWdO909/fc/Ux3L3P3MuA5YIa75/5UiiIigdWrV4cdQsb6TejufgSYA6wHWoGH3b3ZzO4xsxnDHaCIyHDLy8vjqaeeivyN4DUfej80f3d2qT6zR3WZubjNh67pc0Uk9oYyJ3xvz8nVRA9K6CIyAoyU+eWV0EVkROtO3nHowor2GQAREemhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQu/TrjjDMws6w8gKy8zhlnnBFyrYjkHk3OJf1qb2/PuUmLhjIdqkjcqYUuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMRHLS//POOMM2tvbs/Z62brMvKSkhHfeeScrryUicrxYJvRcnHsENP+IiAwvdbmIiMSEErqISEwooYuIxMSAErqZXWlmO8yszcy+mWb/18ysxcxeNLP/a2bnZT9UERHpS78J3czygfuBWmAqMNPMph5X7DdAlbt/BHgE+F/ZDlQkDnLx7k+6A1R8DGSUy8eANnd/FcDMVgHXAC3dBdy9Kan8c8DN2QxSJC40AkuG00ASeinwetL6G8DH+yg/G1iXboeZ3QbcBnDuuecOMEQJm985Fu46PewwUvidY8MOQSTnZHUcupndDFQBn0y3392XAksBqqqqcq+ZImnZ3ftyrlVpZvhdYUchklsGktB3A+ckrU8MtqUws8uBeuCT7t6ZnfBERGSgBjLK5QVgiplNMrNC4AZgTXIBM/tT4IfADHf/ffbDFBGR/vSb0N39CDAHWA+0Ag+7e7OZ3WNmM4Ji3wNGAz8zs61mtqaXlxMRyRqNGko1oD50d38CeOK4bd9JWr48y3GJiPRLo4ZSxXJyrlwclQEamSEiwyuWCT0XR2WARmaIyPDSXC4iIjGhhC4iEhNK6CIiMaGELiISE0roIiIxoYQuIhITSugiIjGhhC4iEhNK6CIiMaGELiISE7G89F8kV2meIRlOSugiJ5HmGZLhpIQuA5JrNxEuKSkJOwSRnKOELv3KZovSzHKyhSoSB7FN6LnWogS1KkVkeMUyoatFKSIjUSwTuoiMDBo1lEoJXUQiS6OGUunCIhGRmFBCFxGJCSV0EZGYUEIXEYkJnRQVOcl0jUR2qT7fp4QuchLpGonsUn2mUpeLiEhMKKGLiMSEErqISEwooYuIxIQSuohITAwooZvZlWa2w8zazOybafYXmdnqYP/zZlaW7UBFRKRv/Q5bNLN84H7gU8AbwAtmtsbdW5KKzQba3f18M7sBWABcPxwBZ9NAx68OpFzUhztlajBjgVWf/cvmsQmqz5Hytz6QcegfA9rc/VUAM1sFXAMkJ/RrgLuC5UeA+8zMPJc/Obn9i4ka1WV2qT6za6TU50C6XEqB15PW3wi2pS3j7keA94A/Ov6FzOw2M9tsZpv37t07tIhFRCStk3pS1N2XunuVu1eNHz/+ZL61iEjsDSSh7wbOSVqfGGxLW8bMTgFOB97ORoAiIjIwA0noLwBTzGySmRUCNwBrjiuzBrglWL4OaMz1/nMRkbjp96Soux8xsznAeiAfWO7uzWZ2D7DZ3dcADwI/MbM24B0SSV9ERE6iAc226O5PAE8ct+07ScsdwGezG5qIiAyGrhQVEYkJJXQRkZiwsM5dmtle4LVQ3nxwzgT+EHYQMaL6zB7VZXZFpT7Pc/e0475DS+hRYWab3b0q7DjiQvWZParL7IpDfarLRUQkJpTQRURiQgm9f0vDDiBmVJ/Zo7rMrsjXp/rQRURiQi10EZGYUEIXEYmJEZvQzWycmX05af0yM3s8zJjiwMx2mdmZGTw/a78HM3vazCIzDM3Mqszs3mF+j03BzzIzu3E43ysTQXzbe9l3j5ld3s/z7zKzr2cxnkgclyM2oQPjgC/3W2qAgmmDYynOny2XuPtmd/9Kpq/T1+/L3auDxTIgZxN6X9z9O+7+VNhx5KIRk9DN7Gtmtj14fBX4LjDZzLaa2feCYqPN7BEz+zczW2HBDQbNbJqZ/dLMtpjZejM7O9j+tJl938w2A39jZp8NXn+bmT0TzicdHDP7n8ENwDeYWUN3qybNZ/ur4AbgvzGzp8xsQlDuj8zsF2bWbGY/Ak64KaOZ5ZvZQ0HdvGRmdwTbzw9ea5uZ/drMJgdP6e338BfB+79kZsvNrKiv7WE7vpVpZl8PWo5Pm9kCM/uVmf3WzP4s2H+ZmT1uZnnBN51xSc992cwmmNl4M/u5mb0QPC4J9t9lZj8xs40kZj6tCF5/q5m9aGZTgnIHgpf8LvBnwf47zOwZM7so6f02mNmFJ6Ga+pJvZsuCY+sXZnZqENtDZnZdsHxVcJxsMbN7j2tFTw3q+lUzO+EfZSyPS3eP/QOYBrwEjAJGA83AnwLbk8pcRuLWeRNJ/KN7FpgOFACbgPFBuetJTCEM8DTw90mv8RJQGiyPC/tzD6Be/gOwFSgGxgAvA1/v5bOV8P6oqC8Bi4Lle4HvBMtXAw6cmab+n0xaHxf8fB74TLBcDJzWx++hmMRtDi8Iyv8j8NXetid9hqoQ67fsuGPs6yTuvft0Uv1dBTyVdAw+Hiz/HfDFYPnjSWVWAtOD5XOB1mD5LmALcGqwvgS4KVguTNp+4Pj3CtZvAb4fLF9AYmrsMI/NMuAIcFGw/jBwc7D8EIn7LnT/7icF2xuS6u8uEn+3RSQu6X8bKIj7cTlSWujTgUfd/aC7HwD+CfizNOV+5e5vuPsxEomuDPgToBJ40sy2Av+DxC+12+qk5Y3AQ2Z2K4m543PdJcA/u3uHu+8H/s9x+5M/20RgvZm9BHwDqAi2Xwr8FMDd1wLtad7nVeCDZrbEzK4E9pnZGBL//B4Nntvh7oeC8r39Hna6+2+DMj8O3ru37bnun4KfW0h8vuOtJtF4gMT9Bbp/F5eTuAn7VhI3lhlrZqODfWvc/d+D5WeBb5vZXBJzf3Rv783PgL80swJgFomkGbad7r41WE5XTx8CXnX3ncF6w3H717p7p7v/Afg9MOG4/bE7LkdKQh+ozqTloyTmizeg2d0vCh4fdvcrksod7F5w978mkfDPAbaY2Qk3yo6Yg0nLS4D73P3DwH8h0QIZEHdvBy4k0TL5a+BH/Twl3e8hio6Q+jeWXGfdn7G3z/cscL6ZjQeu5f1/AHnAJ5KOx9KgkQKpx+JKYAbw78ATZvbnfQUaJK0ngWuAzwErBvD5hlumx0Gfz4/jcTlSEvq/Atea2WlmNgr4DInW9JgBPHcHMN7MLgYwswIzq0hX0Mwmu/vznrj5x15S78WaizYCf2VmxUEr7y/7KHs6799L9pak7c8QnFwzs1oSXTMpLDHqJc/df07iH95Hg28Eb5jZtUGZIjM7rY/33wGUmdn5wfrngV/2sT0XvAX8sSXOMxTRd/2m8MR380eBxSS6Vbrv0fsLoK67XHK/dzIz+yCJ1uu9wD8DHzmuyH5OPP5/RKIL7YUg2eW6HSRa2GXB+vW9Fz1RHI/LEZHQ3f3XJL5C/opE/9iP3H0LsDE4IfK9Pp7bRaK/boGZbSPxVau6l+LfC06AbCfRf7ctix8j69z9BRJf218E1pE4B/BeL8XvAn5mZltInWL0buBSM2sG/iPwuzTPLQWeDroJfgp8K9j+eeArZvYiifo6q49YO4AvBjG8BBwDHuhtez8f/aRw98PAPSSOuyeBfxvkS6wGbia16+srQFVworOFRMsync8B24M6ryTRh5vsReBocOLvjiDeLcA+4B8GGWcogm6kLwP/EhyX++n9+E0ndselLv0f4cxstLsfCFohzwC3Bf8AZYQxsw+Q6H74UNBPnPOSjl8D7gdedvf/HXZcYTGLm8oAAABRSURBVBkRLXTp09KghfJr4OdK5iOTmf1nEt9e66OSzAO3BsdvM4luwR+GHE+o1EIXEYkJtdBFRGJCCV1EJCaU0EVEYkIJXUQkJpTQRURi4v8D+opOn94c2qAAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Use boxplot to visualize credit limit grouped by education level\n",
        "\n",
        "data = []\n",
        "for i in [0,1,2,3]:\n",
        "    temp = df.loc[df.EDUCATION == i, \"LIMIT_BAL\"]\n",
        "    data.append(temp)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "ax.boxplot(data)\n",
        "ax.set_xticklabels([\"others\",\"grad school\", \"university\", \"high school\"])\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MjA6tXAEAmz7"
      },
      "source": [
        "#### (4) Does marital status have anything to do with default risk? Note the credit limit includes the family's total credit.\n",
        "There is no difference of default risk in terms of marital status, although the 'other'marital status group had high default percentage."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "id": "txVSn0SJAmz7",
        "outputId": "36cb2c35-c0a4-4b6d-9c46-3876311d7b84"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Plot a stacked bar chart to visualize default and non-default distribution by marital status \n",
        "\n",
        "create_stacked_bar('MARRIAGE')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1Wgmb5-aAmz8"
      },
      "source": [
        "#### (5) Do younger people tend to miss the payment deadline?\n",
        "\n",
        "Customers aged between 30-50 had the lowest delayed payment rate, while younger groups (20-30) and older groups (50-70) all had higher delayed payment rates. However, the delayed rate dropped slightly again in customers older than 70 years.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "PlSeVvqDAmz8"
      },
      "outputs": [],
      "source": [
        "# Segment the 'AGE' column to 6 groups\n",
        "\n",
        "bins= [21,30,40,50,60,70,80]\n",
        "labels = ['20-30','30-40','40-50','50-60','60-70','70-80']\n",
        "df['AGE'] = pd.cut(df['AGE'],bins=bins, labels=labels,right=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 369
        },
        "id": "7OY1P1ybAmz8",
        "outputId": "4c410b08-a94f-4635-81d4-15659991ca5e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl8AAAFgCAYAAACFexz4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3debxV8/7H8denOg00aFRJMpQGjTJEEqlrSBe/ZFYuQjLPXIS4rnm+JGQoEknJEBIJUanQnAal0jwPp3M+vz/WOqfdcYZdnb12Z/d+Ph77sff6rmF/9vdx9jqf/f1+13eZuyMiIiIi0SiW7ABERERE9iRKvkREREQipORLREREJEJKvkREREQipORLREREJEJKvkREREQipORLRJLOzMqb2YdmttbM3Mx6FtJxPXzU2YF96mTtVxgxiIjkpORLRHaamc0NE5VMM1sXLr9rZkft4KGuBDoBy4FngImFHixgZqPCeLsl4vhxvP+/YhLCd/PY5jwz+8bMVpvZJjObZ2bvmFnTcH3bmGPEPhJSZyJS+EokOwARSQnDgaXAscDZwJlmdr67D4pz/3rh8xvufk8iAtxNXBjz+nQzK+/ua7IKzOwZ4JpwcTQwDagBnAJ8C0yK2X8L8L+Y5YUJiVhECp1avkSkMLzi7v8CGgHvEPywe9HM9gIws9ph681CM1tlZiPM7LBwXT/g0vA4d4etOG3N7EIzmxJ2RW4xsxlm1iPrDc2sX7htr3A53+5CMxsFHB8uvha7b17CGBaY2VIze8TMipvZ/maWEbZMlQ63qxG2/mWX5XKsWuH7bwV+BUoDnWPWH8W2xKunu7dx9+7ufjpQC/goxyE3uvv1MY9H8/ssIrL7UPIlIoXG3bcC94WLlYBjwwRsJNAFmAwMBdoCI82sCjACmBruMxZ4GlgAHAD8DrwFDCRIQJ4zs1Y7Gd57bGsd+jx8nx8K2Oe+ML4ywC1AD3f/IywrT9BVCnA6YMAH7r4pj2OdT3DO/Rp4PSy7KGZ91rEWAy/E7ujua919bo7jlTGzp2Ie5xfwWURkN6HkS0QK27yY19WA04CDgT+B6cAKYD5QFejs7gOAH8PtPw1bcWYBjwL9CJKRZcAfBAnOCTsTlLs/B8wKFweE7/NpAbudFbbo3RUuXxw+9w2fs7oRsxKnAfkcK2vbIcDg8PXxZrZ/+Lpa+DzPw5vumtkPMWO6RuU4XknguphHhwI+i4jsJjTmS0QK2wExr/8CmoWv9yNIEmIdks9xhpF7QlE1j+2LxxXdjslqkZsWPtcKn4cSjHE72cxqA+2AJcCXuR3EzJoAjcPFIe6+wMwmA02AC4CHCeoKYP+YXd8BNgNtcjnsanffZ4c/kYgknVq+RKTQmFkJ4N5wcQUwBpgbLo8Hirm7ubsBFYEH8zjOPmxLvNoQnKs+yVodPq8Pn8uHz4fFEWJG+Bzvua9B+Fw/fF4A4O7pwBtAGkErWGlgoLtn/O0IgdiB9n+E49Ka5Fg3LHyumXU1prs/BXwQZ6wiUkSo5UtECsOlZtaJ4GrHegSDyq909w1m9jEwBzgcGBO2+NQmGPd1KjAql+OtB9YBZYFewBqC1qVYP4fPXc1sK0ELUkH+CJ+vC1ujXnP3SflsP9jMviYYrwbwZsy6vsBNQPtwOdcuRzMrBpwXLv4ILApfFyMYK9bIzJq7+w9m9gLQg+CCgPMJukmPzSO2Mmb2VMzyRne/I5/PIiK7CbV8iUhhOA04BygFvAscmzXNhLuvB04E3iZIuroChxIMpJ+e28HClqWuBGPDjgZWEQyYj/UmQcKTBnQEnowjzscJBv03JOgCrVvA9vcStMBtCvd9PibGaQQtewCz3X1sHsdoS9BduQk41d3PCB+d2JZ4Xhge82rgEoILAY4BuhEkoAOA+3McN+eYr6sK+CwispuwcFyniIjsIDO7jWC81gMpPj+ZiBQiJV8iIjsonLPrPIIuwv2AQ9x9fnKjEpGiQt2OIiI77hDgEWAv4HIlXiKyI9TyJSIiIhIhtXyJiIiIREjJl4iIiEiEitQ8X1WqVPE6deokOwwRERGRAo0fP36Zu//trhxFKvmqU6cO48aNS3YYIiIiIgUys3m5lavbUURERCRCSr5EREREIqTkS0RERCRCRWrMl4iIiOQvPT2dBQsWsGnTpmSHsscoXbo0tWrVIi0tLa7tlXyJiIikkAULFlCuXDnq1KmDmSU7nJTn7ixfvpwFCxZw4IEHxrWPuh1FRERSyKZNm6hcubISr4iYGZUrV96hlkYlXyIiIilGiVe0drS+lXyJiIhIoTIzbrrppuzlxx57jF69eu308erUqcOyZcvy3WbQoEE0aNCAE044Yafeo23bttlziT700EM7dYx4JWzMl5l9BrQAnnb33jnWlQZeAWoD84FL3V0jA0VERApZnduHF+rx5j58WoHblCpVisGDB3PHHXdQpUqVQn3/vLzyyiu8/PLLtG7depeP9dBDD3HnnXcWQlS5S2TL16XALXms6wZMc/fjgOnhsoiIiKSAEiVK0L17d5588sm/rZs7dy4nnngiTZo0oV27dsyfP/9v2yxfvpwOHTrQqFEjLrvsMtw9e91bb73FkUceSbNmzbjiiivIyMjg/vvv59tvv+XSSy/llltuYe7cuRx33HG0aNGCFi1a8N133wEwatQoOnbsmH2snj170q9fv+3e+/bbb2fjxo00a9aMCy64oJBqZHsJS77cfUE+q48HPgpfDwuXRUREJEVcffXV9O/fn9WrV29Xfs0119C1a1cmT57MBRdcwLXXXvu3fe+77z5at27Nb7/9xplnnpmdoE2dOpWBAwcyZswYJk6cSPHixenfvz/33HMPLVu2pH///jz66KNUq1aNzz//nAkTJjBw4MBc3yMvDz/8MGXKlGHixIn0799/1yohD8maaqIysDJ8vQqolNeGZtYd6A5Qu3btxEfWq0Li3yMReq0ueJvdkeo7WkW1vkF1HjXVd7SKan3/+XPu5Z5J+XWzufjMDjzT+w7KlCkF6zfAnz/z/ZhvGfzc3fDnz1zU7jBuvfmmvx3nmy8/Y3Dfx+DPnzmteU0q7lMeFk/myw8/Y/xPP3BEs8MA2LhpM9XKZMCfjWHLOlg6Hf4sTvqatfS8679MnDKD4sWKMeP3+cF7LJsJm5Jf18lKvlYA+4SvK4TLuXL3PkAfgJYtW3pe24mIiMju5frLzqfFyedzyTmdCuV47tD17NP5zx3X5Lvdky/3Z9+qlZn0+TtkZmZS+qBWAJQoUZzMmC7MZE1Em6yrHb8GTg1fnxoui4iISAqpVLECXU5vzytvf5hddkzLJrzz4WcA9B/8Cccd1fxv+7U5ugUDPvgUgE9GjmHlqjUAtGt9JO999AV/LQvabFasXM28BX/+bf/Va9ZRo1oVihUrxpvvDycjIwOAA/arwZQZv7N582ZWrVrFl19+mWvcaWlppKen78Inz1/Cki8ze5lgwH03MxtiZs3MLGsAfj+gsZmNBhqHyyIiIpJibrriIpatWJW9/GzvW3lt4FCanNSFN98fztP33/y3fe69oTvfjJ1AoxM6M/iTkdTerzoADesdRO9be9DhvB40OakL7c+7ikVL/j4FRY+uXXj9vWE0Pekcps2ay957lQFg//2q0+X09hx22GF06dKF5s3/nvgBdO/enSZNmiRswL3FXkGwu2vZsqVnzcGRMBovEC3Vd7SKan2D6jxqqu9oFWJ9T506lQYNGhTa8fKV15iv3V3N3JOuXZFbvZvZeHdvmXNbTbIqIiIiEiElXyIiIiIRUvIlIiIiEiElXyIiIiIRUvIlIiIiEiElXyIiIiIRUvIlIiIihaps3WO3W+43cCg973q4UN+j1+Mv8tiLb+S7zdLlKzmq48U073Aeo8dO2OH36NevHz179gRgyJAhTJkyZadizSlZtxcSERGRKBT23GdFaA64L7/9kcb1D6HvY/fs8rGGDBlCx44dadiw4S4fSy1fIiIiEplhI77Obo066ZwrWbJ0OQBffz+eZu3PpVn7c2ne4TzWrlv/t30ffLov9VqfQesz/sX02fOyy2fP/YOTL7iaw08+n+PO/BfTZs1h4q/TubX303z42SiatT+XjRs3cdXtD9HylAto1KgR9957b/b+derUYdmyYKb8cePG0bZt2+3e97vvvmPo0KHccsstNGvWjNmzZ+9SHajlS0RERArVxk2badb+3OzlFavW0KlDGwBaH9mcH4a9jpnRd8AHPPLC6zx+74089uIbPP/Q7Rx7RDPWrd9A6VIltzvm+MlTeGfoCCZ+/jZbt2bQ4uTzObxJMKN891t78+LDd1H3oNqMnfALPe74DyMH9eH+m69k3OQpPPfg7QA8eNvVVKpYgYx9m9CuXTsmT55MkyZNCvw8xxxzDJ06daJjx4507tx5l+tHyZeIiIgUqjKlSzHx83eyl/sNHMq4ycF4qQWLlnDOVbez6K9lbNmSzoG1awJw7BHNuPG+J7jgzFM465QTqVVz3+2OOXrsz5x58gnsVSa4T2On9scDsG79Br4bP5mzr7g1e9vNW7bkGte7wz6nT//BbLU0Fi1axJQpU+JKvgqbki8RERGJzDV3P8KN3S+kU4fjGfXdOHo98RIAt/e8hNPatebjkWM49oxL+GzA89Q/5MACj5eZmck+5cttl+zlZs78hTz20hv8NPwtKjY8nm7durFp0yYASpQoQWZmJkB2WSJpzJeIiIhEZvWadexXvSoArw8all0+e+4fNG5Ql9uu7sYRTRsxbdbc7fZrc3QLhnz2FRs3bmLtuvUM+/wbAMqXK8uB+9dk0LDPAXB3Jv0242/vu2btevYuU4YK5cuyZMkSPvnkk+x1derUYfz48QC8//77ucZdrlw51q5du/MfPIaSLxEREYlMr5uu4OwrbuPwk8+nSqV9ssuf6juAw048myYndSEtrQSnnLD9dBUtGjfgnNM70LT9uZxy4TUc0axR9rr+zz3IK+8MoelJ59DohM58OGLU3963aaN6ND+sPvXbnMX555/PscduO/69997LddddR8uWLSlevHiucZ977rk8+uijNG/efJcH3Ju779IBotSyZUsfN25cYt+ksC/JjUoRuvR3O6rvaBXV+gbVedRU39EqxPqeOnUqDRo0KLTj5evPn6N5n8JWs3mhHzK3ejez8e7eMue2avkSERERiZCSLxEREZEIKfkSERERiZCSLxERkRRTlMZzp4IdrW8lXyIiIimkdOnSLF++XAlYRNyd5cuXU7p06bj30SSrIiIiKaRWrVosWLCApUuXJv7NVv2V+PdIhNVTC/VwpUuXplatWnFvr+RLREQkhaSlpXHggQXPDF8oeh0dzfsUtiRPpaJuRxEREZEIKfmSoqfqoXDlaLhrEdz6O7R/ICjv9lHwaybrcfRVQflZfeDOhcFzlks/h1pHRB+7iIjs8ZR8SdFTojRMegf6tIXfPoBjr4UD2wTrfvsAnmgQPCa8Afs2goZnwBv/hEZnBstNz4PlM2HBT0n9GEVGXsnuXpWg82tw21y4bR4c3SMoV7IrIpIvjfmSomfRpOAB8PvXcMRlUKZisFy3PdQ5DhaOh49ugK2bITMdls2CjHQolgbH3QivnZq8+IuarGR31hdwZPcg2Z31ORxzLdRoAgMvgq0boey+25LdfqfCJZ/AmKehehMluyIiMZR8SdFVqjwcfyssnw0zR0CxErDyniAJ6Pwa/OMhGNQVvnsWug6F756BRmfA+NehzS1QrQHM+Rq+eSzZn2T3lluyW656kOiO7A1zR2/btvIhSnZFRAqgbkcpmkqVh4s+gL0qw1v/B+kb4df3YeEEmP4J/DUlaIUBGPUwvNQGJg2Eg0+EZTPg4BPgrbOg9Q1Q+eDkfpaiIjbZXb0wKKvfEW6eEXRLHtQWls/KO9ntOgza3JzMTyAiu4O8hjIccw3cNA3uWgyXf7XtHJ6CQxmUfEnRU6ocXDwkSJoGXw4Zm4Oy056AGk2DBKtaA1jy2/b7/eNBGPFvsGJBIlGtERQvCcVLJedzFCU5k921i4LyzWug/9lBHZ75YlCmZFdE8pPbuN0Gp0OH3rB0Orx2ClRvDG3vSNlxu+p2lKKnRlPY7/DgdbfhwfOo/wT/1LsOC5KrOaPh09u37XPwiUE32JxvoHgaLP4FLvkYfhkUtJJJ3rKS3UoHwcALg2R389rgJJm5FbZuAs8IxtfFykp20/ZSsrujSu8Dl30OFfYP6nXmZzD0GqjZPPiRUWE/mP0VDL02qP/zBwatAD/1hS96Bf/crhwNb3WGVfOS/WlEtpfbUAYrHvytblgRtK5nboUt61J23K6SLyl65n4LvSrs2D6zRwYPCL7A/TsXflypKq9k9/1L4fSnofuo4MQ47PJt+yjZ3TWZ6fDlA7DkVzj8kqBlYObnwTjGed/B0J5BS+SJd8Hvo4JunHe7wkWDYfQTwTQrv76vxEt2b9uN2/0s+AHd7h5o2AnWL4WRD8LqP1Jy3K6SLxHJX37J7ssn5l6uZHfXbFkPU4cGr1f/EbRulakIZavBb4ODsY1//AiHngozRgRjHpfPDOq6XPWge6ZP26R+hCIlr5bGDr2hSZdg/ZCrYOKAYPuz+kD902DacBjcPSi79HP47M4i3RUWqdihDK+dChVqBYnXb0Pghxfg3P5w2mMw4JxgKMOoh6HigdDl9eBCn8O7wv+OCaa6+e2DIIErQpR8FYa8vrgXvh/0W1vx4B/YB1fCplX64opIwWq3Cv45pZWBWV8G5w6AzevC57VBMvb7yKDrpsub8Mmt0PZ2+PaJoFWyfM2gtXHCG8n7HEVBbi2NUz+Cxb/CxpVw/G3bttV0Krsut6EMJfcK1mVsCYcyZEK5mtvvl0JDGZR8FYa8vrh/TYVP74Tqh8EZ/4NWPWDKh/riikjB/vwZXjoOGv4TTrw7mE4Fgn9cWc/r/gJ3GH5jUFa7VXAFatpesHeV4IKU63+BX96D9A3J+RxFQW4tjctnBV3kB7XdftsUHYMUqbyGMnz5ABx5WdDtuHxW0CCRJcWGMij5Kgx5fXFjyyDoNtAXV0QKUr1x0B2zci5syUqaHNYvg8POgtXzYf8jYfK72/YxC7rJhlwJB50Ae1eDag2D+e+KFU/GpyhacrY05jVeLr/pVIrwGKRI5TeUYXQedZdiQxk01URhqd0qmJvk1Edh7pjtv7gdegcJ2fh+mgdJRAq2d1Xo9Cz0/AlaXw8/9oGfXoF3L4YqdeGiIUFX48je2/Zp0TUYjL9sZtAakLEZurwRJAKb1ybvsxQVWS2NIx+AQ9pB8wvz3lbTqcguUstXYcnZRdD8Qhj7EnR8Mhiw+e7F2+adSsHBgyJSiGaPhKca/7183hh4oVXu+4zvt+31xpXQ96SEhJaScmtpTN8YDAIvv1+wXLZ6cM5eOWfbfik0BkmipeSrMOT1xT39aWhxMQy/OZjTZK/KsGH5tv30xRURSb69qwbn63LVg8T1xz7BlY0XDwnuFQtw0r3Qshs81SRYTrExSBItJV+FIa8vbqdng/UdnwCeCO6B169jUKYvrojI7iGvlsas83Ve+6TQGCSJlpKvwpDXFze/iUD1xRUREdkjacC9iIiISISUfImIiIhESMmXiIiISIQSmnyZWTcz+87MxphZixzrDjKzb8xslJl9ZWa1EhmLiIiIyO4gYcmXmVUErgXaAhcCz+TYpAfwiru3BV4HrklULCIiIiK7i0S2fB0JjHb3Le4+ByhnZrETWP0G7BO+rgj8lcBYRERERHYLiZxqojKwMmZ5FVAJWBQufwF8ZmaXAqUIkjURERGRlJbIlq8VbGvZAqgQlmX5L/Bvd28C9AIeyu0gZtbdzMaZ2bilS5cmKlYRERGRSCQy+RoLtDazNDOrDaxz980x6w1YFr7+i6BV7G/cvY+7t3T3llWrVk1guCIiIiKJl7BuR3dfaWYvAF8DDlxnZs2A9u7+KNAbeMnMtgJpwBWJikVERERkd5HQ2wu5+6vAqzmKJ4brfgNaJ/L9RURERHY3mmRVREREJEJKvkREREQipORLREREJEJKvkREREQipORLREREJEJKvkREREQipORLREREJEJKvkREREQipORLREREJEIJneFepCB1Ng1Idgg7ZW6yAxARkSJLLV8iIiIiEVLLl8gepKi2NIJaG0UkdajlS0RERCRCSr5EREREIlRg8mVmZczsUTObaGbHmtkzZtYliuBEREREUk08LV9PATcAjYFSQHHglkQGJSIiIpKq4km+zgIejVkeDxyamHBEREREUls8VztmAhaz3BRYl5hwREREpKgoqldQz03y+8eTfA0HbgxfvwlUB/omLCIRERGRFBZP8nU9QcvXaUAa8Doa8yUiIiKyUwpMvtx9DXBJBLGIiIiIpLwCky8zG5lL8Srgc3f/X+GHJCIiIpK64ul2bJtH+T/NrIq7P1CI8YiIpBQNSBaRnOKZauJBYBhQj2CKiWHAk8AAoGviQhMRERFJPfG0fF0N/MfdZwGY2WjgJuBCoHMCYxMREdkhammUoiCe5Gsh8KCZnQ440AqYClQGlicwNhEREZGUE0/ydT7B9BKtw+WfgW5AJeDaxIQlIiIikprimWriF6CFmZUPl9ckPKokUpO1iIiIJFI8U00YcC7BjbVLB4u4u9+U4NhEREREUk483Y7PA1cSjPfKusejEwy6FxEREZEdEM9UE2cSTCsBcB3wFaC5vURERER2QjzJV0VgdPh6EfAe0D1hEYmIiIiksHi6HReH2y0G+gIlgZQedC8iIiKSKPG0fP0bmA3cCGwCVhN0P4qIiIjIDoqn5asYMNXd5wEDzawywW2GRERERGQHxdPy9RpwZMxye7aNARMRERGRHZBny5eZdQLOIJheooeZnRKuak7Q/SgiIiIiOyi/bsfmBLcRcuD48JGlfwJjEhEREUlZ+SVffYDhwI/AXcAIgkRspbvPiSA2ERERkZSTZ/Ll7osI5vUqZmYlgH2B4gBmVtvd50cTooiIiEjqiOfejtcADwOlY4o9nn1FREREZHvxJFD3EQyw/wbYmthwRERERFJbPMnXfOAld/9fooMRERERSXXxJF9TgLvNrCawMixzd38ycWGJiIiIpKZ4kq9zw+e7YsocKDD5MrNuBDfhduAad5+QY/1tBJO2lgDud/eRccQjIiIiUmTFk3xdsjMHNrOKwLXA0cB+wJtA65j1pwAV3P2knTm+iIiISFFUYPLl7q+bWUmgITDH3VfHeewjgdHuvgWYY2blzKyUu28O13cBVprZl8CfQM8dOLaIiIhIkVTgvR3NrDkwGxgHHGFmU83s5TiOXZltY8QAVgGVYpZrApnu3g4YC9wRd9QiIiIiRVQ8N9Z+DlhPcI/HTOAtIJ6uwhXAPjHLFcKy2PWfhq8/BZrkdhAz625m48xs3NKlS+N4WxEREZHdVzzJV1OgX8zyn0C1OPYbC7Q2szQzqw2si+lyBBgFtAxftwRm5XYQd+/j7i3dvWXVqlXjeFsRERGR3Vc8A+4XsO2m2k2A84C5Be3k7ivN7AXga4KrHa8zs2ZAe3d/lCChe9nMvgLSgYt3OHoRERGRIiae5OsRoG/4+nGC7sdu8Rzc3V8FXs1RPDFctxklXCIiIrKHiedqx1fNbDZwWlg03N2/TmxYIiIiIqkpnhtrHwWku/ut4fIxZnaUu49NeHQiIiIiKSaeAfeD2DbmC6BNWCYiIiIiOyie5KsywRxdWdaw/XxdIiIiIhKneAbcTwP+bWZOMNj+LmBqQqMSERERSVHxJF93AR8AzxMkX5uBfyUyKBEREZFUFU/y9R1wBMFYL4AR7p7rhKgiIiIikr98ky8zM4IZ7W909xeiCUlEREQkdeU74N7dHfiYbbcBEhEREZFdEE+3YzPgYDP7J7AkLHN3b5q4sERERERSUzzJ1yHhc9XwISIiIiI7KZ7bC8UzF5iIiIiIxCGuxMrMTjOzp8ysoZldaGbqchQRERHZCQUmX2Z2PTAMuAaoDpwFPJrguERERERSUjwtX9ez/b0cvwBaJCYcERERkdQWT/JVEZgUs7wXUDwx4YiIiIiktniudhwLXBW+vhloDYxJWEQiIiIiKSyelq9rgY0E93U8GVhE0BUpIiIiIjsonqkmpplZA+DQsGi6u2ckNiwRERGR1FRg8mVmFYC7gbZh0Vdm1tvdVycyMBEREZFUFM+YrwHAKUBmuNwCaAB0TFRQIiIiIqkqnjFfxwB9gDIEVzr2DctEREREZAfF0/L1EbDU3dMBzGwpMDShUYmIiIikqHiSryOB883s4nC5FjDDzCYD7u661ZCIiIhInOJJvuqGz/vHlB2a24YiIiIikr94ppqI6+bbIiIiIlIwJVYiIiIiEVLyJSIiIhIhJV8iIiIiESow+TKzi83sgJjlymameb5EREREdkI8LV+vEUw3kaU9MDox4YiIiIiktjyvdjSzTsAZgAE9zOyUcFVzYFMEsYmIiIiknPymmmgOdAMcOD58ZOmfwJhEREREUlZ+yVcfYDjwI3AXMIIgEVvp7nMiiE1EREQk5eSXfG0Gfgeq5FxhZpXcfUXCohIRERFJUfklX8sIWrpy4wXsKyIiIiK5yC+B+oa8ky8RERER2Ql5Jl/u3jbCOERERET2CAV2HZrZPbkUu7s/kIB4RERERFJaPOO2euVS5oCSLxEREZEdFE/y1YVtY78qAlcDPyUsIhEREZEUVmDy5e7vxS6b2XrU6iUiIiKyU+IZ8zU5x/YHAmsSFpGIiIhICoun2/GwmNeZwFzgtoREIyIiIpLiihW0gbsXi3mUcPdD3P39eA5uZt3M7DszG2NmLfLY5j4zm7WjgYuIiIgURXHNUm9mrYDGQOmsMnd/poB9KgLXAkcD+wFvAq1zbLMvUG/HQhYREREpuuKd5+veXFblm3wBRwKj3X0LMMfMyplZKXffHLPN3cB/gMHxBiwiIiJSlBXY7QhcDnwavv4P8AvwWBz7VQZWxiyvAiplLZhZXaCsu0/OuWMsM+tuZuPMbNzSpUvjeFsRERGR3Vc8yVc14KPw9VjgBaBzHPutAPaJWa4QlmXpRRxTVrh7H3dv6e4tq1atGsfbioiIiOy+4uDMva0AABmZSURBVBnztYzgKsdVwLNAqfBRkLFAbzNLA2oA63J0OR4EPG9mADXM7Bl3v3ZHgpc9U+fDa/HY2U23K3t33B8UM+h8+P7ZZa9+O4f7P5rCbScfykWt6jBh3koue30cWzIyee785nz262KGTV4UdfgiIrKHiyf5egpYDdwHPEmQiN1S0E7uvtLMXgC+Jpgh/zozawa0d/dH3b1V1rZmNkuJl8Tro8l/8u3MZQD847Dq3NepEd/OXEabelWYMG8lPfpPAGD95q2UK1WCq9oewoV9x/J4l6a0qVeFtZu2UnnvUkq8REQkKfJMvszsfmAgsAT43t3nmVlfwNx9fTwHd/dXgVdzFE/MZbtD4g9Z9nSb0jNZnL4JgFMOq87StZv45NdFtKlXhYY1y/Pxdccxc8la7v7wV/5YsZH0jEx+X7qOjVsy2Jrh3HVaA24ZlO9QQ4mRV0vjre9NxgyG9WzNYftV4Lb3JzPwpz+4uNUB3PyPQ5mzdD2X9PuJFeu3cPsp9Vm9IZ3/fT07SZ9CRGT3kd+Yr7uA5sBrBFcu4u4b4k28RBKtbrWyHH1QZQb8+AfpGc6nvy7mwr5j6dF/PAdXK8ujnZuyMT2DXkN/46WLWvLtrGXUqrQX4+etpH2jfXn78qPp1alRsj/Gbu+jyX9y9ENfcvRDX3Lv0N8AslseO7eoRd1qZbfb/sb29bj3w98oZkanpjWpU3kv2h5alb7f/h557CIiu6P8uh2XA88DBjxnZv+NWefufnBCIxMpwMWt6pCekcmAsfMA+GLqX9nrfvh9Oe0b7AtA/7Hz6T92PuVLl+C9q47h1vcm89ZlR9HsvhGMuKENx9WtwugwmZC/y6ulsUxacW7qcCivfDuHHidsa7zesjWTP1ZsYO3mdDZvzeDfHRvy8MfTSM/wZH0EEdmN5NWanla8GMfXq0qZtOJMWrCKmwdNYsHKjSk5bje/lq+HgKwB8uWBqjGPagmOSyRfe5cszhnNa/L5lCUsWRP8md55agNa1N6H5vvvw1EHVmLa4rXb7XND+3q8PPp3Nm7JoGTxYtSvUY6ypUpQskQ8F/1KzpbGK48/iKmL1vDNzO2ngPn3kF+5u2NDlq3dzMoN6bg7+1YoTf/LjuKJLk0pnab6FtmT5dWavmT1Ji7t9xNXvjWelgdU5Ib29bLH7V755ngOrV6ONvWqcNSBlYr8uN08W77c/SngKTP7Crjf3b+KLiyR/J3VohblSqfx5vfzsssq7Z3Gq92OoFSJ4kxeuIq7Pvg1e90h1crSeL8K3DdsCgBDJy3k3Sta8eOcFXwzQ/PHxSO2pbHS3iW5pPWBdHnxeyrtXRKAYhZsN2LKEkZMWUKJYsaHPY/lpncnMbRna459eCSPd2nKmc334+0f/0jiJxGRZMqrNX3opD8BMIMNWzLYp0waWzM9JcftFni1o7ufYGZNzOx6YCjBrYLmurvOnpI0b/4wjzd/mLdd2c35fBln/bWOzi9+v922+W0v28vZ0tiwRnnKl07j0+vbZG/zn7OasGpDOp/8uhiArsfU4Yupf7Fg5UZKFDPq1yhH5bIlKVmieLI+RpGRV7dMn29+59HOTTikWll+nr+KW96bxJI1m3nynGZ0aLgvI6Ys4YaBwTVNg686ht7DpzBh/qpkfASRAmW1pj/95czthiVc364uZUuV4PXv5uY7brdXp0ZMX7KWXmHrWVFSYPu/mZ0DTAAeB+oA9xDM9yUie4icLY2zl67j9Ge/5fRnv+XOwb8A8MyXMxkzOxg7V2nvknQ+vBb/GzWLdZu38vyoWfzvwsPZsjWTDycuTNrnKCry6pZ59rzmZGQ6XV76njqV96b3GY2pX70cpzauzoV9x3Ja4xrUr16O/2uxH78vW6fEawc0rFGeQVe2Ysr9/+CHO9rR8oCKHFe3Cl/ceDyTe3XghQtaUL5M0F5x28mH8ut9/+CNfx1JyeLBv9Hnzm/O6U1qJPMjFDk5x+0C9Gh7MNecWJd7PvyVb8KxuP3Hzuf0577lkU+ncXGrA/hw4p9cefzBXPTKWNrUrcJxdask6yPstHjm+bof+BI4KVz+CLgzYREVUQ1rlOe+fzaiUc3yrNm4lZ4DJtC8dkUuO+5AKpRJY/ritdz2/mSmLV6rX6lS5ORsady8NZNfFq4G4JeFqxnw4/zttl+xfgunPD06e/nxETN4fMSMaIJNAbl1y/y6cDUNapTnweFTmbpoLV/PWMp5R+7Pfz6eytYMZ/aydWzNzKREcaNH20M4p8/3BbyLZNm7ZHFe/9eRzF2+nrNe+I5aFctQrJjx3PktGPv7cm4YOJH+lx/Fbf+oz8OfTNPcgYUgt3G7PdoezK0n1+eZL2fyxdS/qFq2FEvXbZubPZXG7cYTcU1gZMzyVqBMYsIpmrK+uABnvfAd/x7yCxvTM7jrtAbMXLKOs1/8noY1y3ND+3r6lSoicYu9yKFCmTQgmDwYYN3mdEoUL8baTVt5efTvDLjsaPp88zunNa7B2z/Np+eJdXn78qPpeaKmUSzIifWrUbVcKR75dBrTFq/li6l/sXFLBhXKpPHNjKX8snA1M5es44T61fIcg1QUu76SKbdxu+cfVRuAa9vV5Yc72/HB1cdkr8satzto3AKmL1mbPW53yqI1RXLcbjwtX78AF4evLwJOBiYlLKIiKOuL26P/eKYtXsu0xWspUcxYsHIDqzZsYe6y9WRkOus2b2XL1kz9ShWRuMR2y6SF3VtlSwen7bKl0sjIdFZs2MJTX8zkqS9mUrvSXrxwQQseHzGD846ozclPf8PEezowfPIi5izTFI152a9i0J5w28n1qVNlb6YuWsPTX8wkM9NpUKM8e5cszv4Vy1CudFrKjkGKWm7jdlv/N+/r+lJt3G48yddNBF2NBnQluDn2zYkMqqjJ7Yt7x+BfGDB2PreeXJ+TD6vO8vVbeGLEDBau2pjnr9RD9y3HmNnLeG7krCR/IhFJtty6ZaYvXkuHhvsyeuZSjq9XlVHT/yIjc9tA5bs7NuDBj6eyV8nilCtdgkOrlyOteLHscUmSu5Ub0gH4+Y9VPPLZdAZcdhQrWm3hqS9ncF27epxzxP6s35LBn6s2Apo7UHZdgd9Id/8eOAToGD7quvvYRAdWlMR+cXv0n0Crgyrz79MacuvJ9Rk+eRHn9PmB4sWM+/8ZzKb+1Bcz6fjstwyesJDj6lZl9l/rOe6QKlz86liuOv5gDqyydzI/jojsBnLrlrnm7QmUKGYMuvIY5q/YsN10KsfVrcKWrc73s5fzzYylTFm0hnevaMXQSQuZvmRtbm8hoW9nLmPL1kzSt2ayOT0DJxjX+NW0pZz2zGiuHvAzG7ZsZdD47S/yT6UxSBKtfFu+zKwK0BNoSXBz7PHATxHEVaTk9sXNuiomPSOTTekZuDvVK5Tebj/9ShWRvOTWLTNjyTrOeOG7XLcfPXNZdmtLeobT7TWdquO1cNVGbh40iZs61KPrMXX4duYyHv1sOvf/sxEnHlqNFRu28MGEhfQdPSd7H80dKLsivxtr1wHGANUJuhwBTgUuM7Nj3H1+HrvucXL74t76/mS6tNyfi1sdwCmHVef3Zevp/dHU7H1if6WmFTf9ShURSaKhk/7MnuQzy1VvTchz+1QbgyTRyq/lqzdQA3gZ+IQgATsZuAx4gGD8l4Ry++I+/9Usnv8q9/Fb+pUqIiKyZ8ov+WoDvO7uV8SUfWBmJYETExuWiIiISGrKb3BRVSC36z5HoRtri4iIiOyU/Fq+SgHXmlnnHOU1gZKJC0lEREQkdRU0z1eL8JGT51ImIiIiIgXIL/k6MLIoRERERPYQeSZf7j4vr3UiIiIisnM0m6eIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiERIyZeIiIhIhJR8iYiIiEQoocmXmXUzs+/MbIyZtcix7lYzGxuue9bMLJGxiIiIiOwOEpZ8mVlF4FqgLXAh8EyOTT5w96Pc/VhgX+DERMUiIiIisrtIZMvXkcBod9/i7nOAcmZWKmulu8+M2XYzsDWBsYiIiIjsFhKZfFUGVsYsrwIq5dzIzI4HagDf5HYQM+tuZuPMbNzSpUsTEqiIiIhIVBKZfK0A9olZrhCWZTOzJsDDwLnu7rkdxN37uHtLd29ZtWrVhAUrIiIiEoVEJl9jgdZmlmZmtYF17r45a6WZHQK8SpB4LUtgHCIiIiK7jYQlX+6+EngB+Bp4G7jezJqZ2S3hJk8RtIy9bmajzOy0RMUiIiIisrsokciDu/urBK1bsSaG6zom8r1FREREdkeaZFVEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQkq+RERERCKk5EtEREQkQglNvsysm5l9Z2ZjzKxFjnWlzay/mY0On0snMhYRERGR3UHCki8zqwhcC7QFLgSeybFJN2Caux8HTA+XRURERFJaIlu+jgRGu/sWd58DlDOzUjHrjwc+Cl8PC5dFREREUloik6/KwMqY5VVApTzW51wnIiIikpJKJPDYK4B9YpYrhGW5rc+5LpuZdQe6h4vrzGx6IccZpSrAskQc2P6biKMWearvaCWsvkF1ngf9jUdL9R2tVKjvA3IrTGTyNRbobWZpQA1gnbtvjln/NXAqMDF8/jq3g7h7H6BPAuOMjJmNc/eWyY5jT6H6jpbqO3qq82ipvqOVyvWdsG5Hd18JvECQVL0NXG9mzczslnCTfkBjMxsNNA6XRURERFJaIlu+cPdXgVdzFE8M120Ezkvk+4uIiIjsbjTJarRSovu0CFF9R0v1HT3VebRU39FK2fo2d092DCIiIiJ7DLV8iYiIiERIyZeIiIhIhJR8iYiIiEQooVc7SsDMKgG4e64TyYoUZfr7jpaZ7QvsFy4udPclyYwn1am+o7cnnFM04D5BzKw28AjQjuD2SQaUB0YCt7v73ORFl3rMrAJwB3AGUA1w4C/gQ+Bhd1+VxPBSjv6+o2dmzYAXCe4IsjAsrkVQ/z3cfUKyYktFqu9o7WnnFCVfCWJm3wNPAe+5e0ZYVhw4G7je3Y9OZnypxsw+I/iSvu7ui8Oy6kBXoJ27d0hmfKlGf9/RM7OJwBXuPjZH+dHAS+7eNDmRpSbVd7T2tHOKkq8EMbOZ7l53R9fJzjGz6e5+6I6uk52jv+/oFVDns9z9kKhjSmWq72jtaecUjflKnPFm9gLwOvBHWLY/QUvMz0mLKnXNM7NbCVq+lkD2WI1ubKt/KTz6+47eJ2Y2HHiD7ev8YuDTpEWVulTf0dqjzilq+UoQMysJXAr8k5jBmsBQ4JUcNxmXXWRmFYHbCeq7Wli8hKC+/5vKAzeTQX/fyWFmp5BLnbv7x8mLKnWpvqOzp51TlHyJiIiIREjzfCWIme1lZrea2S1mVtrMuprZUDN7xMzKJju+PYGZjUx2DKnKzKrkWL7QzJ4xs+5mZsmKK5WZWc+sejezg83sGzNbaWZjzaxxsuNLNWY22Mwu0Pk6GmZ2kJm9amYPmFlZM3vZzH41s0FmVifZ8RU2JV+J0w/YFzgQGA4cATxKcPns/5IXVmoys8k5Hr8Ax2YtJzu+FDQi64WZ/Ru4CBgPtAeeSFZQKe4qd18Wvn4GeNLdKwK3EUyJIIXrKOBMYL6ZvWtmZ4ZdY5IY/YCfgPXAD8B04BSC8XWvJi+sxFC3Y4KY2UR3bxa2AiwCari7h8uT3L1JkkNMKWY2FFgD9AY2EiS5o4HWAO4+L3nRpR4z+9ndm4evJwDHuft6M0sDJri7WmIKWexVu2b2k7sfEbNuss4phSvrb9zMyhOMQzqP4Ef0R8Db7j4i3wPIDslxTpnv7rVzW5cq1PKVYB5ktx+Hz1nLyngLmbt3At4H+gBNwwn50t19nhKvhChjZs3N7HCguLuvB3D3dCAjuaGlrPfMrJ+ZHQR8YGbXm9kBZnYJMD/ZwaWgrHP2Gnd/091PBeoDYwku7pHClWlm9czsCGAvM2sJYGaHAMWTG1rhU8tXgphZX4KJ4dblKD+YYDqE1smJLLWZ2d7AA8DBwOHuXivJIaUkM/sqR9H57r7IzCoDn7l7y2TElerMrBtwFcHfdymCS/KHEFzRuzqJoaUcM/vG3dskO449hZm1A14AMoHLgRuApgSz3F/u7h8mMbxCp+QrCczMXBWfUGbWFGjl7hoLE6FwRupS7r4h2bGISNEWXmCyMmvG+1SibscImVkfyO56lARy90lA9WTHsScxs17unqHEKzpZ5xSJhuo7cg+lYuIFSr6ipq6YaHVKdgB7GNV39HROiZbqO1opW99KvqL1V7ID2MNovqloqb6jp3NKtFTf0UrZ+taYL0lZZlbM3TOTHceeQvUtIhIftXwliJlVMLOHzWyama0ws+VmNjUs2yfZ8aUaMythZleY2acxE6sON7Mrw7mnpBCpvqOnc0q0VN/R2tPqW8lX4rwLrATaunsld68MnBCWvZvUyFLTm0AzoBdwavi4j+BS5beSF1bKUn1HT+eUaKm+o7VH1be6HRMkdjbqHVknO8fMZrh7vR1dJztH9R09nVOipfqO1p5W32r5Spx5FtxYe9+sAjPb18xuI5gYUQrXCjM728yy/6bNrJiZnUPwy0kKl+o7ejqnREv1Ha09qr6VfCXOOUBl4GszW2lmK4BRQCWgSzIDS1HnAp2BxWY2w8xmAIuBs8J1Uriy6ntJWN8zUX0nms4p0VJ9R2uPqm91OyaQmdUHagE/xN5myMxOdvdPkxdZajKzowjuxzab4B5srYAp7v5xUgNLceEthQCedvcLkxrMHsTMjgOOBH7RTZ4LX3g+mebuq81sL4L7ObYAfiOY/FO3cypEZnYt8IG7p1wrV26UfCVI+Id0NTCVYGDydVn3pjKzCe7eIpnxpRozuxc4BSgBfE7wT2kU0J7gXoMPJi+61GNmQ3MpPhEYCdk3OpdCZGY/uvuR4evLCM4vQ4AOwDB3fziZ8aUaM/sNaOruW8OZ7dcD7wPtwvKzkhpgijGz1QR1PBsYAAxy92XJjSpxlHwliJn9QnBvwXVmVgd4D3jT3Z82s5/dvXlSA0wxYX03I7jZ8GKglruvMbMywFh3b5LUAFOMmU0ApgB9CVobDXibsMvR3b9OXnSpKfa8YWY/Aae6+1ILbib/g7s3Tm6EqcXMprp7g/D1dj+YzWyiuzdLXnSpx8x+Bg4HTiLoguwEjCc4rwx297VJDK/QacxX4hTL6mp097lAW+AUM3sCzQSeCFtj7is4293XALj7RkATfxa+lgQnxruA1e4+Ctjo7l8r8UqYYmZWMezmNXdfCuDu64GtyQ0tJf1qZpeEryeZWUsAM6sHpCcvrJTl7p7p7iPc/VKgJvACcDLwe3JDK3xKvhJniZll/zIKE7GOQBVAv1AL35ZwXAYEv56AYOI+lHwVuvAk+SRwCXCXmT1H0OUriVOBIOEdB1QysxoAZlYW/aBLhMuA481sNtAQ+N7MfgdeDtdJ4drub9jd0919qLufBxyQpJgSRt2OCWJmtQhaYxbnsu5Ydx+ThLBSlpmVcvfNuZRXAWq4+y9JCGuPYWanAce6+53JjmVPE/7o2Nfd5yQ7llRkZuWBAwl+XCxw9yVJDiklmVk9d5+R7DiiouRLREREJELqdhQRERGJkJIvERERkQgp+RKRlGVmV5uZh49DY8oPMrN3zGyZmW0yszlm1jdcVydmn6zHquR9ChFJNbo6SURS2TkEV7sWI7hFyQNmVh34HqgK9APGAAfx99si/Qw8Er7eEkWwIrJnUPIlIinJzGoCxwLvAm0IErEHCGaGrwb0cfcrYrbvneMQS4Evwtea10lECo26HUUkVZ1NcI4bBAwGGplZQ7bNA/cJgJmVC6ck2dvM0mL270CQgC0FPowsahFJeUq+RCRVdSHoLpwGjA3LziG4HRIxz2+yLck6Nmb/sQT3Bm0P3JToYEVkz6FuRxFJOWa2P9CKYNbs32JWdSFoCTuVIKn6ELgnXPfPHIdZ5u5fICJSyJR8iUgqOpsg8foP8GNYdinBLb6+BK4ArjKzYgQtXJVyOUZNM4sdhP++u2vsl4jsMs1wLyIpx8zGAkcQ3HZnaVjWBRgIPAi8BvwXaAeUAhYAnwP3A2WA3G7VU9HdNeWEiOwyJV8iIiIiEdKAexEREZEIKfkSERERiZCSLxEREZEIKfkSERERiZCSLxEREZEIKfkSERERiZCSLxEREZEIKfkSERERidD/A0zfYo0bVLLxAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Plot a stacked bar chart to visualize default and non-default by age\n",
        "\n",
        "create_stacked_bar('AGE')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8sEs1su8Amz9"
      },
      "source": [
        "#### (6) Has the repayment status changed in the 6 month from April 2005 (PAY_6) to September 2005(PAY_0)? \n",
        "There was a huge jump from May,2005 (PAY_5) to July, 2005 (PAY_3) when delayed payment increased significantly, then it peaked at August, 2005 (PAY_2), things started to get better in September, 2005 (PAY_1). "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "TMc-vkJ1Amz9",
        "outputId": "a92b1033-4673-4b1d-ca85-77f1fa437544"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Subset a dataframe with the records that have default\n",
        "\n",
        "has_default = df[df['HAS_DEF']== 1]\n",
        "default_trend = has_default[['PAY_6','PAY_5','PAY_4','PAY_3','PAY_2','PAY_1']].sum(axis=0)\n",
        "\n",
        "# Draw a line chart to show the trend. The lower the number, the shorter delayed payment\n",
        "fig,ax = plt.subplots()\n",
        "ax.plot(default_trend)\n",
        "plt.xticks(['PAY_6','PAY_5','PAY_4','PAY_3','PAY_2','PAY_1'],['Apr','May','Jun','Jul','Aug','Sep'])\n",
        "\n",
        "plt.xlabel('Months in 2005',fontweight='bold')\n",
        "plt.ylabel('Total delayed months',fontweight='bold')\n",
        "plt.title('Delayed payment trend',fontweight='bold')\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1SZYUHAQAmz-"
      },
      "source": [
        "#### (7) Is there any correlation between bill amount and the delayed time?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QJ2NEqhFAmz-"
      },
      "source": [
        "There isn't an obvious correlation between bill amount and the total months of delay."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 820
        },
        "id": "8yaDyGhTAmz-",
        "outputId": "444d2b54-2440-43f8-e232-273c38f92733"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 6 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Make 6 boxplots to visualize bill amounts and the total months of delayed payment\n",
        "\n",
        "from matplotlib.pyplot import figure\n",
        "\n",
        "# Make \"PAY_1\" to\"PAY_6\" , 'BILL_AMT1\" to'BILL_AMT6\" into a list respectively\n",
        "pay_list = [ f\"PAY_{i}\" for i in range(1, 7) ]\n",
        "bill_amt_list = [ f\"BILL_AMT{i}\" for i in range(1, 7) ]\n",
        "\n",
        "fig, ax = plt.subplots(3,2, figsize=(10, 10))\n",
        "\n",
        "for pay, bill_amt, myax in zip(pay_list, bill_amt_list, ax.flatten()):\n",
        "    \n",
        "    # Prepare data for boxploting\n",
        "    data = []\n",
        "    for i in sorted(has_default[pay].unique()):\n",
        "        temp = has_default.loc[has_default[pay] == i, bill_amt]\n",
        "        data.append(temp)\n",
        "        \n",
        "    # Make boxplot for each PAY variable\n",
        "    myax.boxplot(data, showfliers=False,) \n",
        "    myax.set_xticklabels(sorted(has_default[pay].unique()))\n",
        "    \n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mnbJrdLUAmz_"
      },
      "source": [
        "#### (8) Is there any correlation between credit limit and the default payment next month?\n",
        "\n",
        "Unsurprisingly, customers who had higher credit limits had lower delayed payment rates."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "ygbxxm4LAmz_",
        "outputId": "a7e77790-504f-48de-a3e4-4aa35c55271a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Make a boxplot to visualize credit limit and default payment next month\n",
        "# 1: default next month; 0: no default next month\n",
        "\n",
        "def0 = df.loc[df['DEF_PAY_NMO'] == 0,'LIMIT_BAL']\n",
        "def1 = df.loc[df['DEF_PAY_NMO'] == 1,'LIMIT_BAL']\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "ax.boxplot([def0, def1],  showfliers=False)\n",
        "\n",
        "ax.set_xticklabels(['No_default',\"Default\"],fontweight ='bold')\n",
        "ax.set_ylabel('Credit limit',fontweight ='bold')\n",
        "ax.set_title('Credit limit & default next month',fontweight ='bold')\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Aca5_6PZAm0A"
      },
      "source": [
        "Customers with high credit limits tend to have higher 'no-default'rate."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tH4dzg4wAm0A"
      },
      "source": [
        "\n",
        "#### (9) Why are there negative bills?\n",
        "\n",
        "Interestingly enough, the minimals of those 6 bill columns are negative numbers. In general, there are 599-688 bills with negative amounts each month, which is less than 2% of total 30,000 records monthly. Also, the average negative amount in USD is $38 - $102 which is a small amount. Therefore, an informed guess would be these are refunds from the last billing cycle. This will be investigated in data analysis."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 36,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "tss-a6VnAm0A",
        "outputId": "ce12508b-3335-4e3e-ddf2-f993f9672178"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           BILL_AMT1      BILL_AMT2     BILL_AMT3      BILL_AMT4  \\\n",
              "count   30000.000000   30000.000000  3.000000e+04   30000.000000   \n",
              "mean    51223.330900   49179.075167  4.701315e+04   43262.948967   \n",
              "std     73635.860576   71173.768783  6.934939e+04   64332.856134   \n",
              "min   -165580.000000  -69777.000000 -1.572640e+05 -170000.000000   \n",
              "25%      3558.750000    2984.750000  2.666250e+03    2326.750000   \n",
              "50%     22381.500000   21200.000000  2.008850e+04   19052.000000   \n",
              "75%     67091.000000   64006.250000  6.016475e+04   54506.000000   \n",
              "max    964511.000000  983931.000000  1.664089e+06  891586.000000   \n",
              "\n",
              "           BILL_AMT5      BILL_AMT6  \n",
              "count   30000.000000   30000.000000  \n",
              "mean    40311.400967   38871.760400  \n",
              "std     60797.155770   59554.107537  \n",
              "min    -81334.000000 -339603.000000  \n",
              "25%      1763.000000    1256.000000  \n",
              "50%     18104.500000   17071.000000  \n",
              "75%     50190.500000   49198.250000  \n",
              "max    927171.000000  961664.000000  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9e73bdb4-1962-4164-b303-45a77d189ea6\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>BILL_AMT1</th>\n",
              "      <th>BILL_AMT2</th>\n",
              "      <th>BILL_AMT3</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>3.000000e+04</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "      <td>30000.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>51223.330900</td>\n",
              "      <td>49179.075167</td>\n",
              "      <td>4.701315e+04</td>\n",
              "      <td>43262.948967</td>\n",
              "      <td>40311.400967</td>\n",
              "      <td>38871.760400</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>73635.860576</td>\n",
              "      <td>71173.768783</td>\n",
              "      <td>6.934939e+04</td>\n",
              "      <td>64332.856134</td>\n",
              "      <td>60797.155770</td>\n",
              "      <td>59554.107537</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-165580.000000</td>\n",
              "      <td>-69777.000000</td>\n",
              "      <td>-1.572640e+05</td>\n",
              "      <td>-170000.000000</td>\n",
              "      <td>-81334.000000</td>\n",
              "      <td>-339603.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>3558.750000</td>\n",
              "      <td>2984.750000</td>\n",
              "      <td>2.666250e+03</td>\n",
              "      <td>2326.750000</td>\n",
              "      <td>1763.000000</td>\n",
              "      <td>1256.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>22381.500000</td>\n",
              "      <td>21200.000000</td>\n",
              "      <td>2.008850e+04</td>\n",
              "      <td>19052.000000</td>\n",
              "      <td>18104.500000</td>\n",
              "      <td>17071.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>67091.000000</td>\n",
              "      <td>64006.250000</td>\n",
              "      <td>6.016475e+04</td>\n",
              "      <td>54506.000000</td>\n",
              "      <td>50190.500000</td>\n",
              "      <td>49198.250000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>964511.000000</td>\n",
              "      <td>983931.000000</td>\n",
              "      <td>1.664089e+06</td>\n",
              "      <td>891586.000000</td>\n",
              "      <td>927171.000000</td>\n",
              "      <td>961664.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9e73bdb4-1962-4164-b303-45a77d189ea6')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9e73bdb4-1962-4164-b303-45a77d189ea6 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9e73bdb4-1962-4164-b303-45a77d189ea6');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ],
      "source": [
        "# Get statistic summary of bill statement columns\n",
        "# The min numbers are negative\n",
        "\n",
        "bill = df[['BILL_AMT1','BILL_AMT2','BILL_AMT3','BILL_AMT4','BILL_AMT5','BILL_AMT6']]\n",
        "bill.describe()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "id": "qYzI7q2JAm0B",
        "outputId": "a9f13b05-67cd-4984-ec84-e0694a8c61bf"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            amount\n",
              "bill_cycle        \n",
              "BILL_AMT1      590\n",
              "BILL_AMT2      669\n",
              "BILL_AMT3      655\n",
              "BILL_AMT4      675\n",
              "BILL_AMT5      655\n",
              "BILL_AMT6      688"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-bcc9bb31-bf0d-47f7-b60c-9424eac048d2\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>amount</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>bill_cycle</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>BILL_AMT1</th>\n",
              "      <td>590</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_AMT2</th>\n",
              "      <td>669</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_AMT3</th>\n",
              "      <td>655</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <td>675</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <td>655</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_AMT6</th>\n",
              "      <td>688</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-bcc9bb31-bf0d-47f7-b60c-9424eac048d2')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-bcc9bb31-bf0d-47f7-b60c-9424eac048d2 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-bcc9bb31-bf0d-47f7-b60c-9424eac048d2');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ],
      "source": [
        "# How many bills have negative amount?\n",
        "# There are 3932 bills with negative amounts, each month has 599-688 records \n",
        "\n",
        "bill_melt = pd.melt(bill, var_name = 'bill_cycle',value_name = 'amount')\n",
        "neg_bill = bill_melt[bill_melt['amount']<0]\n",
        "neg_bill.groupby('bill_cycle').count()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JbqAOHESAm0B",
        "outputId": "fb35e0f2-f585-487e-bcf4-f48025366c40"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Average negative amounts in NTD are: \n",
            "\n",
            "bill_cycle\n",
            "BILL_AMT1   -1154.796610\n",
            "BILL_AMT2   -1230.621824\n",
            "BILL_AMT3   -1466.291603\n",
            "BILL_AMT4   -1631.242963\n",
            "BILL_AMT5   -1634.708397\n",
            "BILL_AMT6   -3074.492733\n",
            "Name: amount, dtype: float64\n",
            "\n",
            "Average negative amounts in USD are: \n",
            "\n",
            "bill_cycle\n",
            "BILL_AMT1    -38.493220\n",
            "BILL_AMT2    -41.020727\n",
            "BILL_AMT3    -48.876387\n",
            "BILL_AMT4    -54.374765\n",
            "BILL_AMT5    -54.490280\n",
            "BILL_AMT6   -102.483091\n",
            "Name: amount, dtype: float64\n"
          ]
        }
      ],
      "source": [
        "# Get the average amount of negative bill each month\n",
        "# Use USD/NTD = 30 to get the amount in USD. The average negative amount is $38-$102\n",
        "\n",
        "neg_bill_mean_ndt = neg_bill.groupby('bill_cycle')['amount'].mean()\n",
        "\n",
        "print('Average negative amounts in NTD are: \\n')\n",
        "print(neg_bill_mean_ndt)\n",
        "print('\\nAverage negative amounts in USD are: \\n')\n",
        "print(neg_bill_mean_ndt/30)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3iIbMv6BAm0C"
      },
      "source": [
        "#### (10) Why are some bill statement amounts greater than credit limit? Could it be late payment interest?\n",
        "\n",
        "The common sense is that the bill statement amount shouldn't exceed credit limit, however, there are 3931 customers whose bill amounts are greater than credit limit. Could the difference be late payment interest assuming these customers had delayed payment? "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wbAVmRUYAm0C",
        "outputId": "6732ae26-e4cc-49c5-eb2d-4089aef569dc"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3931, 25)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "# Is there any bill amount that is greater than credit limit?\n",
        "\n",
        "condition1 = df['BILL_AMT1'] > df['LIMIT_BAL'] \n",
        "condition2 = df['BILL_AMT2'] > df['LIMIT_BAL'] \n",
        "condition3 = df['BILL_AMT3'] > df['LIMIT_BAL'] \n",
        "condition4 = df['BILL_AMT4'] > df['LIMIT_BAL'] \n",
        "condition5 = df['BILL_AMT5'] > df['LIMIT_BAL'] \n",
        "condition6 = df['BILL_AMT6'] > df['LIMIT_BAL'] \n",
        "\n",
        "large_bill = df[condition1 | condition2 |condition3 | condition4 | condition5 | condition6]\n",
        "large_bill.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n1tog6x9Am0D",
        "outputId": "056a59d8-e436-48a2-f785-272c1c0a28bb"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.0    2184\n",
              "1.0    1747\n",
              "Name: HAS_DEF, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ],
      "source": [
        "large_bill[\"HAS_DEF\"].value_counts()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "miBOb5JKAm0D"
      },
      "source": [
        "#### (11) Why did customers who had no consumption in 6 months then default in the next month?\n",
        "\n",
        "First explore if there is any customer who never had consumption and the data shows there are 870 customers who were inactive in 6 months. Then check against the 'default.payment.next.month' column to see if these customers all had no default. The common sense should be no default payment next month if the previous bill was 0. However, we see there are 317 \"1\"s which means default in the next month. We will investigate this in the data analysis process."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 455
        },
        "id": "rgfRS1-QAm0D",
        "outputId": "c7d9f7ea-790f-43b6-db8f-7afcb149eae1"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       BILL_AMT1  BILL_AMT2  BILL_AMT3  BILL_AMT4  BILL_AMT5  BILL_AMT6\n",
              "ID                                                                     \n",
              "19           0.0        0.0        0.0        0.0        0.0        0.0\n",
              "20           0.0        0.0        0.0        0.0        0.0        0.0\n",
              "46           0.0        0.0        0.0        0.0        0.0        0.0\n",
              "80           0.0        0.0        0.0        0.0        0.0        0.0\n",
              "122          0.0        0.0        0.0        0.0        0.0        0.0\n",
              "...          ...        ...        ...        ...        ...        ...\n",
              "29909        0.0        0.0        0.0        0.0        0.0        0.0\n",
              "29910        0.0        0.0        0.0        0.0        0.0        0.0\n",
              "29974        0.0        0.0        0.0        0.0        0.0        0.0\n",
              "29980        0.0        0.0        0.0        0.0        0.0        0.0\n",
              "29986        0.0        0.0        0.0        0.0        0.0        0.0\n",
              "\n",
              "[870 rows x 6 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9c06a17a-bf67-4845-8428-f09617dfd9f8\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>BILL_AMT1</th>\n",
              "      <th>BILL_AMT2</th>\n",
              "      <th>BILL_AMT3</th>\n",
              "      <th>BILL_AMT4</th>\n",
              "      <th>BILL_AMT5</th>\n",
              "      <th>BILL_AMT6</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ID</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29909</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29910</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29974</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29980</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29986</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>870 rows × 6 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9c06a17a-bf67-4845-8428-f09617dfd9f8')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9c06a17a-bf67-4845-8428-f09617dfd9f8 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9c06a17a-bf67-4845-8428-f09617dfd9f8');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ],
      "source": [
        "# Is there any customer who never had any transaction, meaning bill amount in 6 months remained as 0\n",
        "# There are 870 customers whose bill amount was 0 in 6 months\n",
        "\n",
        "bill_amt = df[['BILL_AMT1','BILL_AMT2','BILL_AMT3','BILL_AMT4','BILL_AMT5','BILL_AMT6']]\n",
        "no_transaction = bill_amt[bill_amt.sum(axis=1) ==0]\n",
        "no_transaction"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 42,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 455
        },
        "id": "A1cJJAuPAm0E",
        "outputId": "b6087e15-c26c-438d-a86b-fc2a1c189083"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       DEF_PAY_NMO\n",
              "ID                \n",
              "19               0\n",
              "20               0\n",
              "46               1\n",
              "80               1\n",
              "122              1\n",
              "...            ...\n",
              "29909            1\n",
              "29910            0\n",
              "29974            1\n",
              "29980            0\n",
              "29986            0\n",
              "\n",
              "[870 rows x 1 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-7f347ed8-5ac3-40de-8bb1-4aca3fe09cd4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>DEF_PAY_NMO</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ID</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29909</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29910</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29974</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29980</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29986</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>870 rows × 1 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7f347ed8-5ac3-40de-8bb1-4aca3fe09cd4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-7f347ed8-5ac3-40de-8bb1-4aca3fe09cd4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-7f347ed8-5ac3-40de-8bb1-4aca3fe09cd4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ],
      "source": [
        "# Check if those 870 customers defaulted in the next month (Oct.2005)\n",
        "# The assumption is since they did not have any consumption, they shouldn't default, so there shouldn't be any '1'\n",
        "\n",
        "no_transaction_de = df.loc[no_transaction.index,['DEF_PAY_NMO']]\n",
        "no_transaction_de"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3d4618JbAm0E",
        "outputId": "3ae55ad1-80b6-473b-aa10-63a5a7ee37e7"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    553\n",
              "1    317\n",
              "Name: DEF_PAY_NMO, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 43
        }
      ],
      "source": [
        "# How many inactive customers had default payment next month\n",
        "# 317 customers had default payment next month which is against common sense\n",
        "\n",
        "no_transaction_de['DEF_PAY_NMO'].value_counts()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W8zsycQWAm0F"
      },
      "source": [
        "## 5. Statistical Analysis"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hjUdh11NAm0F"
      },
      "source": [
        "Should this section be embeded to previous EDA step?\n",
        "- Correlation between credit limit and bill amount: Pearson R \n",
        "- Correlation between credit limit and payment default: t-test\n",
        "- Correlation between education and  payment default: chi-squared test\n",
        "- Correlation between age and payment default: chi-squared test\n",
        "- Correlation between sex and payment default: permutation test\n",
        "- Correlation between sex and credit limit: t-test\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6VwFNYYYAm0F"
      },
      "source": [
        "#### Correlation between credit limit and bill amount.\n",
        "\n",
        "We assume customers who have a high credit limit will have a higher bill amount. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 44,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 112
        },
        "id": "T7Xxb_beAm0G",
        "outputId": "71982ce0-acc1-4605-de01-8e19bc78b677"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           LIMIT_BAL  BILL_SUM\n",
              "LIMIT_BAL   1.000000  0.302044\n",
              "BILL_SUM    0.302044  1.000000"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-24fd9b90-d2ed-48a0-bcd9-214430f55fad\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <th>BILL_SUM</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>LIMIT_BAL</th>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.302044</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BILL_SUM</th>\n",
              "      <td>0.302044</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-24fd9b90-d2ed-48a0-bcd9-214430f55fad')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-24fd9b90-d2ed-48a0-bcd9-214430f55fad button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-24fd9b90-d2ed-48a0-bcd9-214430f55fad');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ],
      "source": [
        "df['BILL_SUM'] = df[['BILL_AMT1','BILL_AMT2','BILL_AMT3','BILL_AMT4','BILL_AMT5','BILL_AMT6']].sum(axis=1)\n",
        "credit_bill = df[['LIMIT_BAL','BILL_SUM']]\n",
        "credit_bill.corr()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k54-Ue_FAm0G",
        "outputId": "19999de2-1af1-4e71-a9b4-d728d0a0b855"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.30204379702615786, 0.0)"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ],
      "source": [
        "from scipy import stats\n",
        "\n",
        "credit_limit = df['LIMIT_BAL']\n",
        "bill_amount = df[['BILL_AMT1','BILL_AMT2','BILL_AMT3','BILL_AMT4','BILL_AMT5','BILL_AMT6']].sum(axis=1)\n",
        "\n",
        "stats.pearsonr(credit_limit, bill_amount)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3CrYfqO1Am0H"
      },
      "source": [
        "The Pearson R shows there is a positive correlation between these 2 variables."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n7fyqdkbAm0H"
      },
      "source": [
        "#### Correlation between credit limit and payment default.\n",
        "Ideally, we should have customers' income data, but since this data is not available, we can assume higher income customers have higher credit limits. Therefore, we will verify if there is a correlation between credit limit and default payment using t-test.\n",
        "\n",
        "- Null hypothesis: credit limit does not affect default likelihood.\n",
        "- Alternative hypothesis: credit limit impact default likelihood.\n",
        "- Set significance level $\\alpha$ to 0.05"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 46,
      "metadata": {
        "id": "zIagsK62Am0H"
      },
      "outputs": [],
      "source": [
        "no_def_limit = df.loc[df['HAS_DEF']==0,'LIMIT_BAL']\n",
        "def_limit = df.loc[df['HAS_DEF']==1,'LIMIT_BAL']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 47,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S1a_FqE6Am0H",
        "outputId": "92ca64f2-6d9b-438b-9c11-ebe1f4fc426e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Ttest_indResult(statistic=46.758765680556394, pvalue=0.0)"
            ]
          },
          "metadata": {},
          "execution_count": 47
        }
      ],
      "source": [
        "from scipy import stats\n",
        "from scipy.stats import ttest_ind_from_stats\n",
        "\n",
        "ttest_ind_from_stats(mean1=np.mean(no_def_limit),std1=np.std(no_def_limit),nobs1=len(no_def_limit),\n",
        "                     mean2=np.mean(def_limit),std2=np.std(def_limit),nobs2=len(def_limit))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H61YsJA6Am0I"
      },
      "source": [
        "We get a p value as 0, therefore we need to reject the null hypothesis and accept the alternative hypothesis. Credit limit has an impact on payment default."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QmsBCEC9Am0I"
      },
      "source": [
        "#### Correlation between education and payment default.\n",
        "\n",
        "Previous visualization indicates education impacts default likelihood. We will use a Chi-squared test to verify this finding.\n",
        "- Null hypothesis: education does not affect default likelihood.\n",
        "- Alternative hypothesis: education impacts default likelihood.\n",
        "- Set significance level $\\alpha$ to 0.05"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 48,
      "metadata": {
        "id": "rDyhvjjJAm0I"
      },
      "outputs": [],
      "source": [
        "# Get the contigence table of education and default data\n",
        "\n",
        "edu_def_contigence = pd.crosstab(df['HAS_DEF'], df['EDUCATION'], margins=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 49,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hbBEa_9HAm0J",
        "outputId": "ca794ed4-d35b-431b-ac98-fbe040f1bc9d"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9.997508691343157e-76"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ],
      "source": [
        "# Use Chi-squared test to test if education affects default likelihood.\n",
        "\n",
        "from scipy.stats import chisquare\n",
        "from scipy.stats import chi2_contingency\n",
        "stat, p, dof, expected = chi2_contingency(edu_def_contigence)\n",
        "\n",
        "p"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EzsvkFxSAm0J"
      },
      "source": [
        "Since the p value is close to 0, we will reject the null hypothesis and accept the alternative hypothesis. Because education has a strong correlation with default probability, we should keep this variable in the machine learning model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GN_Wb03tAm0J"
      },
      "source": [
        "#### Correlation between age and payment default.\n",
        "\n",
        "Previous visualization indicates age impacts default likelihood. We will use a chi-squared test to verify this finding.\n",
        "- Null hypothesis: age does not affect default likelihood.\n",
        "- Alternative hypothesis: age impacts default likelihood.\n",
        "- Set significance level $\\alpha$ to 0.05"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 50,
      "metadata": {
        "id": "0S4XcW8rAm0K"
      },
      "outputs": [],
      "source": [
        "# Get the contigence table of age and default data\n",
        "\n",
        "age_def_contigence = pd.crosstab(df['HAS_DEF'], df['AGE'], margins=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 51,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fQtwhk-wAm0K",
        "outputId": "84ca8697-5c73-47aa-c7d5-623fbf71e377"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6.185243361338736e-26"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ],
      "source": [
        "# Use chi-squared test to test if age affects default likelihood.\n",
        "\n",
        "from scipy.stats import chisquare\n",
        "from scipy.stats import chi2_contingency\n",
        "stat, p, dof, expected = chi2_contingency(age_def_contigence)\n",
        "\n",
        "p"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qK7TBBrKAm0K"
      },
      "source": [
        "The p value is smaller than significance level $\\alpha$, we will reject the null hypothesis and accept the alternative hypothesis, which is age has impact on default probability."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TWq_LNC7Am0L"
      },
      "source": [
        "#### Correlation between sex and payment default. \n",
        "\n",
        "In previous data visualization, it appears males tend to default more than females. Does sex have any correlations with default or was this observation due to chance event? Let's find out with a permutation test on each group's default proportions and mean respectively.\n",
        "- Null hypothesis: sex has no impact on default probability. \n",
        "- Alternative hypothesis: sex has impact on default probability. \n",
        "- Set significance level $\\alpha$ to 0.05."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 52,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EFb6aKUHAm0L",
        "outputId": "11b72e49-2a32-433e-fd59-5b081992e015"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.03849496062187008"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ],
      "source": [
        "# Get the true difference of 'HAS_DEF' proportions of two sex groups\n",
        "\n",
        "male = df.loc[df['SEX']=='M','HAS_DEF']\n",
        "female = df.loc[df['SEX']=='F','HAS_DEF']\n",
        "true_diff_prop = male.sum()/len(male) - female.sum()/len(female)\n",
        "true_diff_prop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "metadata": {
        "id": "7TfdJ01YAm0L"
      },
      "outputs": [],
      "source": [
        "# Define functions to do permutation test\n",
        "\n",
        "def permutation_sample(data1, data2):\n",
        "    data = np.concatenate((data1,data2))\n",
        "    permuted_data = np.random.permutation(data)\n",
        "    perm_sample_1 = permuted_data[:len(data1)]\n",
        "    perm_sample_2 = permuted_data[len(data1):]\n",
        "\n",
        "    return perm_sample_1, perm_sample_2\n",
        "\n",
        "def diff_of_prop(data1,data2):\n",
        "    diff = data1.sum()/len(data1) - data2.sum()/len(data2)\n",
        "\n",
        "    return diff\n",
        "\n",
        "def draw_perm_reps(data1, data2, func, size=1):\n",
        "    perm_replicates = np.empty(size)\n",
        "\n",
        "    for i in range(size):\n",
        "        perm_sample_1, perm_sample_2 = permutation_sample(data1,data2)\n",
        "        perm_replicates[i] = func(perm_sample_1,perm_sample_2)\n",
        "\n",
        "    return perm_replicates"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "id": "8fx43u1RAm0M"
      },
      "outputs": [],
      "source": [
        "# Draw 10000 permutation tests\n",
        "\n",
        "perm_replicates = draw_perm_reps(male,female,diff_of_prop,size=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 55,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nA0rUfhyAm0M",
        "outputId": "3e7a31b9-3e75-408c-e597-391daf0e142d"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.0"
            ]
          },
          "metadata": {},
          "execution_count": 55
        }
      ],
      "source": [
        "# Calculate p value - the probability of getting a mean difference equal or greater than the observed mean difference\n",
        "\n",
        "p = np.sum(perm_replicates >= true_diff_prop) / len(perm_replicates)\n",
        "p"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QIdivSvRAm0M"
      },
      "source": [
        "Test conclusion: since the p value is 0, we should reject the null hypothsis and accept the alternative hypothsis, which is sex has impact on default likelihood. To be more concrete in this case, male customers tend to default more."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jQxVQ_NSAm0N"
      },
      "source": [
        "#### Correlation between credit limit and sex.\n",
        "\n",
        "Since we know the credit limit is strongly correlated with default probability. Let's see if sex plays any role in credit limit and default likelihood. \n",
        "- Null hypothesis: sex has no impact on credit limit. \n",
        "- Alternative hypothesis: sex has impact on credit limit. \n",
        "- Set significance level $\\alpha$ to 0.05."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rcLRCB87Am0N",
        "outputId": "51913e48-4de6-4cea-cc45-5e7bbb547838"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "163519.8250336474 170086.46201413427\n"
          ]
        }
      ],
      "source": [
        "# Calculate the true mean difference of credit limit\n",
        "\n",
        "male_limit= df.loc[df['SEX']=='M','LIMIT_BAL']\n",
        "female_limit = df.loc[df['SEX']=='F','LIMIT_BAL']\n",
        "print(np.mean(male_limit),np.mean(female_limit))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DhIre9AWAm0N",
        "outputId": "2ce9afbb-1233-4ed8-dfac-0b41603b6791"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Ttest_indResult(statistic=-4.289049447941391, pvalue=1.8000013013822418e-05)"
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ],
      "source": [
        "# Use t-test to verify if different sex groups have the same mean of credit limit.\n",
        "\n",
        "from scipy import stats\n",
        "from scipy.stats import ttest_ind_from_stats\n",
        "\n",
        "ttest_ind_from_stats(mean1=np.mean(male_limit),std1=np.std(male_limit),nobs1=len(male_limit),\n",
        "                     mean2=np.mean(female_limit),std2=np.std(female_limit),nobs2=len(female_limit))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UBem_WLQAm0O"
      },
      "source": [
        "The p value is much lower than  𝛼 , so we will reject null hypothesis and accept alternative hypothesis. We have verified that sex plays a role in credit limit and it is not due to chance."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#2. Machine Learning Modeling"
      ],
      "metadata": {
        "id": "sAMkpahfBitI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd \n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "%matplotlib inline \n",
        "import warnings\n",
        "warnings.simplefilter(\"ignore\")\n",
        "from pprint import pprint\n",
        "import joblib\n",
        "import imblearn\n",
        "\n",
        "from sklearn.model_selection import KFold\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.model_selection import StratifiedShuffleSplit\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import cross_validate\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "from sklearn import metrics  \n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import plot_confusion_matrix\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import plot_roc_curve\n",
        "from sklearn.metrics import precision_recall_curve\n",
        "from sklearn.metrics import plot_precision_recall_curve\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "import xgboost as xgb\n",
        "from xgboost import XGBClassifier"
      ],
      "metadata": {
        "id": "DqFNm8z7Bym8"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data = pd.read_csv(\"/content/UCI_Credit_Card_Processed.csv\")\n",
        "data.columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VMQJTayeBzRY",
        "outputId": "c2b6b88e-9cc4-440e-ec94-03b0f3e63b2d"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['ID', 'LIMIT_BAL', 'SEX', 'EDUCATION', 'MARRIAGE', 'AGE', 'PAY_1',\n",
              "       'PAY_2', 'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1', 'BILL_AMT2',\n",
              "       'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT1',\n",
              "       'PAY_AMT2', 'PAY_AMT3', 'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6',\n",
              "       'DEF_PAY_NMO'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "##Prepare for Modeling\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "1. Feature Engineering"
      ],
      "metadata": {
        "id": "KVfcNjwUCGd_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Bin ‘AGE’ data to 6 groups\n",
        "bins= [21,30,40,50,60,70,80]\n",
        "labels = list(range(6))\n",
        "data['AGE'] = pd.cut(data['AGE'],bins=bins, labels=labels,right=False)\n",
        "\n",
        "# Covert categorical column into integer by extracting the code of the categories\n",
        "data.AGE = data.AGE.cat.codes"
      ],
      "metadata": {
        "id": "1FZOiwXHCA0P"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "2. Feature Selection\n",
        "\n",
        "---\n"
      ],
      "metadata": {
        "id": "4NWxE8V5CVTg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define predictor variables and target variable\n",
        "X = data.drop(columns=['ID','DEF_PAY_NMO'])\n",
        "y = data['DEF_PAY_NMO']\n",
        "\n",
        "# Save all feature names as list\n",
        "feature_cols = X.columns.tolist() \n",
        "\n",
        "# Extract numerical columns and save as a list for rescaling\n",
        "X_num = X.drop(columns=['SEX', 'EDUCATION', 'MARRIAGE', 'AGE'])\n",
        "num_cols = X_num.columns.tolist()"
      ],
      "metadata": {
        "id": "55VVjCagCQ1c"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "3. Check Class Imbalance\n",
        "\n",
        "---\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "SUPCfeuYCi_5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Check target classes balance\n",
        "cla_bal = data['DEF_PAY_NMO'].value_counts(normalize=True)\n",
        "print(cla_bal)\n",
        "\n",
        "# Plot the classes\n",
        "cla_bal.plot(kind = 'bar')\n",
        "plt.title('Nondefault(0) and default(1) comparison',fontweight = \"bold\")\n",
        "plt.xlabel('Classes')\n",
        "plt.ylabel('Percentage')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 344
        },
        "id": "wSVk07O4Cd8E",
        "outputId": "0f3e5282-5637-443b-ec17-98ee05f670f9"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0    0.7788\n",
            "1    0.2212\n",
            "Name: DEF_PAY_NMO, dtype: float64\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "With typical default classification problems, we expect imbalanced classes as we know most people will not default. This dataset is also imbalanced, with 78% non-default vs. 22% default.\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "\n",
        "##4. Split Training and Test Data"
      ],
      "metadata": {
        "id": "iD9bYtVYCwrS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define function to split data with and without SMOTE \n",
        "\n",
        "def data_split(X, y, imbalance = False):\n",
        "    # split data \n",
        "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,shuffle=True, stratify=y, random_state=42)\n",
        "    \n",
        "    if imbalance:\n",
        "        # use SMOTE to over sample data\n",
        "        from imblearn.over_sampling import SMOTE\n",
        "        sm = SMOTE(random_state = 42)\n",
        "        X_train, y_train = sm.fit_resample(X_train, y_train.ravel())\n",
        "    \n",
        "    return X_train, X_test, y_train, y_test"
      ],
      "metadata": {
        "id": "H1obM8JUCrd3"
      },
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "5. Rescale Features\n",
        "\n",
        "---\n",
        "\n"
      ],
      "metadata": {
        "id": "CYuk9e1tC_oG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define function to rescale training data using StandardScaler\n",
        "\n",
        "def standard_scaler(X_train, X_test, numerical_cols):\n",
        "    \n",
        "    # Make copies of dataset\n",
        "    X_train_std = X_train.copy()\n",
        "    X_test_std = X_test.copy()\n",
        "    \n",
        "    # Apply standardization on numerical features only\n",
        "    for i in numerical_cols:\n",
        "        scl = StandardScaler().fit(X_train_std[[i]])     # fit on training data columns\n",
        "        X_train_std[i] = scl.transform(X_train_std[[i]]) # transform the training data columns\n",
        "        X_test_std[i] = scl.transform(X_test_std[[i]])   # transform the testing data columns\n",
        "    \n",
        "    return X_train_std,X_test_std"
      ],
      "metadata": {
        "id": "_7oP80BeC7Q5"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Classification Predictive Modeling\n"
      ],
      "metadata": {
        "id": "VC09ue5PDQiz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "##1. Logistic Regression\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "Rescaling the features variables then run Logistic Regression with and without SMOTE sampling to compare ROC_AUC scores."
      ],
      "metadata": {
        "id": "BsI6H0cjDWwN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def run_logistic_regression(imbalance=False):\n",
        "    # Split data\n",
        "    X_train, X_test, y_train, y_test = data_split(X, y, imbalance = imbalance)\n",
        "\n",
        "    # Rescale data\n",
        "    X_train_std, X_test_std = standard_scaler(X_train, X_test, numerical_cols = num_cols)\n",
        "    \n",
        "    # Instantiate model\n",
        "    clf_lr = LogisticRegression(random_state=42)\n",
        "    \n",
        "    # Fit the model\n",
        "    clf_lr.fit(X_train_std, y_train)\n",
        "\n",
        "    # Use model's default parameters to get cross validation score\n",
        "    scores = cross_val_score(clf_lr, X_train_std, y_train, scoring =\"roc_auc\", cv = 5)\n",
        "    roc_auc_lr = np.mean(scores)\n",
        "    \n",
        "    if imbalance:\n",
        "        return \"Logistic Regression\", \"With SMOTE\", roc_auc_lr\n",
        "    else:\n",
        "        return \"Logistic Regression\", \"Without SMOTE\", roc_auc_lr\n",
        "\n",
        "model_result = []\n",
        "model_result.append(run_logistic_regression())\n",
        "model_result.append(run_logistic_regression(imbalance=True))\n",
        "\n",
        "pd.DataFrame(model_result, columns = [\"Model\", \"SMOTE\", \"ROC_AUC Score\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 112
        },
        "id": "YOTxS2PFDJ98",
        "outputId": "99111e17-71f2-4142-9df8-e46fd239005d"
      },
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 Model          SMOTE  ROC_AUC Score\n",
              "0  Logistic Regression  Without SMOTE       0.794684\n",
              "1  Logistic Regression     With SMOTE       0.794689"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-658e8a4d-3076-4bcb-ba81-9d036aaffbae\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>SMOTE</th>\n",
              "      <th>ROC_AUC Score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>Without SMOTE</td>\n",
              "      <td>0.794684</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>With SMOTE</td>\n",
              "      <td>0.794689</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-658e8a4d-3076-4bcb-ba81-9d036aaffbae')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-658e8a4d-3076-4bcb-ba81-9d036aaffbae button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-658e8a4d-3076-4bcb-ba81-9d036aaffbae');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 76
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Since SMOTE sampling yields a much better ROC_AUC score on training data, we will use SMOTE while tuning the model.\n",
        "\n",
        "---\n",
        "\n"
      ],
      "metadata": {
        "id": "3Iumm2RKHtqN"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Tune Parameters of Logistic Regression"
      ],
      "metadata": {
        "id": "mYPWc70dH0NO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Randomized search for the best C parameter\n",
        "\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "from scipy.stats import uniform\n",
        "\n",
        "# Split data with SMOTE \n",
        "X_train, X_test, y_train, y_test = data_split(X, y, imbalance = True) \n",
        "\n",
        "# Rescale data\n",
        "X_train_std, X_test_std = standard_scaler(X_train, X_test, numerical_cols = num_cols)\n",
        "\n",
        "logistic = LogisticRegression(solver='saga', tol=1e-2, max_iter=200,random_state=42)\n",
        "distributions = dict(C=uniform(loc=0, scale=4), penalty=['l2', 'l1'])\n",
        "clf = RandomizedSearchCV(logistic, distributions, random_state=42)\n",
        "\n",
        "lr_best= clf.fit(X_train_std, y_train)   \n",
        "\n",
        "print(lr_best.best_params_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hgowq2hHDihI",
        "outputId": "42b0689b-d42c-44ce-c07a-73c272be1c3f"
      },
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'C': 2.832290311184182, 'penalty': 'l1'}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save tuned model and parameters\n",
        "#joblib.dump(lr_best,\"logreg_sm_tuned.pkl\") "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_Bt4kZvCH56A",
        "outputId": "955280dc-96b5-4328-e296-ad1d04e3ccbd"
      },
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['logreg_sm_tuned.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "lr_best = joblib.load(\"logreg_sm_tuned.pkl\")"
      ],
      "metadata": {
        "id": "JnLs-jk5IPoc"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get ROC_AUC score of tuned model on training data\n",
        "\n",
        "scores_tuned = cross_val_score(lr_best, X_train_std, y_train, scoring = \"roc_auc\", cv = 5)\n",
        "roc_auc_lr_best = np.mean(scores_tuned)\n",
        "\n",
        "print(f'ROC_AUC score after tuning parameters:{roc_auc_lr_best:.3f}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jao7gC20IQ3M",
        "outputId": "86e4e02c-ad40-45ea-b791-a31b07ba6b6f"
      },
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "ROC_AUC score after tuning parameters:0.793\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save best ROC_AUC \n",
        "\n",
        "joblib.dump(roc_auc_lr_best,\"logreg_ROC_AUC_CV.pkl\") "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rrRU17LkIWbN",
        "outputId": "b89982a9-f5d4-44bf-d633-7ed9c76f2b78"
      },
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['logreg_ROC_AUC_CV.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "roc_auc_lr_best = joblib.load(\"logreg_ROC_AUC_CV.pkl\")\n",
        "roc_auc_lr_best"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hBX-iuMTIfZk",
        "outputId": "583a5080-2343-45be-f204-42e7bad1a860"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7926199602390442"
            ]
          },
          "metadata": {},
          "execution_count": 82
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "After fine-tuning parameter C, Logistic Regression model got ROC_AUC training score as 0.792, slightly decreased from original score 0.797. This could mean model is less overfitting the training data."
      ],
      "metadata": {
        "id": "fcF_jYrRIkmy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Logistic Regression Model Evaluation"
      ],
      "metadata": {
        "id": "oPao2GwnJYoh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Precision_Recall and F1 Score"
      ],
      "metadata": {
        "id": "clESVN8BJf5P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define a function to compute Precision, Recall and F1 score\n",
        "\n",
        "def get_pre_rec_f1(model,X_test,y_test):\n",
        "    y_pred = model.predict(X_test)\n",
        "    tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
        "    \n",
        "    precision = tp / (tp + fp)\n",
        "    recall = tp / (tp + fn)\n",
        "    F1 = 2 * (precision * recall) / (precision + recall)\n",
        "    \n",
        "    print(f'Precision:{precision:.3f}\\nRecall:{recall:.3f}\\nF1 score:{F1:.3f}')"
      ],
      "metadata": {
        "id": "mgsz_fSpIm3D"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate precision and recall of Logistic Regression model\n",
        "\n",
        "print('Logistic Regression model on test data:')\n",
        "get_pre_rec_f1(lr_best, X_test_std, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9f6wuEp7IkLO",
        "outputId": "9d702b75-2969-409c-9050-ce1ca9ad7d39"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Logistic Regression model on test data:\n",
            "Precision:0.729\n",
            "Recall:0.714\n",
            "F1 score:0.721\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model Performance:\n",
        "\n",
        "The business nature of credit card default prediction requires model to have a high recall. Logstic Regression model has a recall as 0.714. Let's look at another model - Random Forest to see if we can get a better recall."
      ],
      "metadata": {
        "id": "1xBo-EoOJ0tT"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##2. Random Forest\n",
        "Tree based model does not need feature to be rescaled, so we will skip rescaling step."
      ],
      "metadata": {
        "id": "3SWKEvBoKFfV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def run_random_forest(imbalance=False):\n",
        "    # Split data  \n",
        "    X_train, X_test, y_train, y_test = data_split(X, y, imbalance = imbalance)\n",
        "    \n",
        "    # Instantiate model\n",
        "    clf_rf = RandomForestClassifier()\n",
        "   \n",
        "    # Fit the model\n",
        "    clf_rf.fit(X_train, y_train)\n",
        "\n",
        "    # Use model's default parameters to get cross validation score\n",
        "    scores = cross_val_score(clf_rf, X_train, y_train, scoring =\"roc_auc\", cv = 5)\n",
        "    roc_auc_rf = np.mean(scores)\n",
        "    \n",
        "    if imbalance:\n",
        "        return \"Random Forest\", \"With SMOTE\", roc_auc_rf\n",
        "    else:\n",
        "        return \"Random Forest\", \"Without SMOTE\", roc_auc_rf\n",
        "\n",
        "model_result = []\n",
        "model_result.append(run_random_forest())\n",
        "model_result.append(run_random_forest(imbalance=True))\n",
        "\n",
        "pd.DataFrame(model_result, columns = [\"Model\", \"SMOTE\", \"ROC_AUC Score\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 112
        },
        "id": "w9pa9I_mIjih",
        "outputId": "0550e924-eff5-4a9f-d03e-50b9ffe61862"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           Model          SMOTE  ROC_AUC Score\n",
              "0  Random Forest  Without SMOTE       0.903371\n",
              "1  Random Forest     With SMOTE       0.903748"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-3c2fd5dc-d9ee-4466-9848-11b81590ea10\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>SMOTE</th>\n",
              "      <th>ROC_AUC Score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Random Forest</td>\n",
              "      <td>Without SMOTE</td>\n",
              "      <td>0.903371</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Random Forest</td>\n",
              "      <td>With SMOTE</td>\n",
              "      <td>0.903748</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3c2fd5dc-d9ee-4466-9848-11b81590ea10')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-3c2fd5dc-d9ee-4466-9848-11b81590ea10 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-3c2fd5dc-d9ee-4466-9848-11b81590ea10');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Again, SMOTE sampling has a better performance. With the default parameters, Random Forest model generated ROC-AUC score 0.916 on training data, which is better than Logistic Regression model. Next, we will try to fine tune the hyperparameters with SMOTE sampling."
      ],
      "metadata": {
        "id": "ip4V_8JRKTCU"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Tune Parameters of Random Forest"
      ],
      "metadata": {
        "id": "5D7vZh-LKeju"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Split data with SMOTE\n",
        "X_train_sm, X_test, y_train_sm, y_test = data_split(X, y, imbalance = True)"
      ],
      "metadata": {
        "id": "sJvOzKqvKLbw"
      },
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create parameter grid  \n",
        "param_grid = {\n",
        "    'max_depth': [60, 90, 110],\n",
        "    'min_samples_leaf': [3, 4, 5],\n",
        "    'min_samples_split': [8, 10, 12],\n",
        "    'n_estimators': [100, 200, 300]\n",
        "}\n",
        "\n",
        "# Instantiate the model\n",
        "clf_rf = RandomForestClassifier()\n",
        "\n",
        "# Instantiate grid search model\n",
        "grid_search = GridSearchCV(estimator = clf_rf, param_grid = param_grid,    \n",
        "                          cv = 3, n_jobs = -1, verbose = 1)\n",
        "\n",
        "# Fit grid search to the data\n",
        "grid_search.fit(X_train_sm, y_train_sm)\n",
        "grid_search.best_params_"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vhKJx5s9KqDO",
        "outputId": "c2d09fcf-38f7-4f65-9363-6efcdb9edce8"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 3 folds for each of 81 candidates, totalling 243 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'max_depth': 90,\n",
              " 'min_samples_leaf': 3,\n",
              " 'min_samples_split': 8,\n",
              " 'n_estimators': 300}"
            ]
          },
          "metadata": {},
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save tuned model\n",
        "\n",
        "#joblib.dump(grid_search,\"RandomForest_tuned_final.pkl\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S3cJYYnJKwkY",
        "outputId": "e57cc026-57d6-4b92-d3e2-e33fe05d8361"
      },
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['RandomForest_tuned_final.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "grid_search = joblib.load(\"RandomForest_tuned_final.pkl\")"
      ],
      "metadata": {
        "id": "o0r9-xeeSTHo"
      },
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Use the best parameters to fit the model\n",
        "\n",
        "rf_best = RandomForestClassifier(**grid_search.best_params_)   \n",
        "rf_best.fit(X_train_sm,y_train_sm)\n",
        "\n",
        "scores_best = cross_val_score(rf_best, X_train_sm, y_train_sm, scoring =\"roc_auc\", cv = 5)\n",
        "roc_auc_best = np.mean(scores_best)\n",
        "\n",
        "print(f'ROC_AUC training score after tuning for Random Forest: {roc_auc_best:.3f}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uEOL1zVeScOe",
        "outputId": "8ef10e45-e45a-4587-a007-8000c4880a61"
      },
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "ROC_AUC training score after tuning for Random Forest: 0.897\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "After tuning parameters, Random Forest model's ROC-AUC score decreased slightly to 0.897.\n",
        "\n",
        "---\n",
        "\n",
        "#### Random Forest Model Evaluation\n",
        "\n",
        "\n",
        "---\n",
        "Precision_Recall and F1 Score\n"
      ],
      "metadata": {
        "id": "8xNi5izPSl0N"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compute precision, recall and F1 score of Random Forest model on test data\n",
        "\n",
        "print('Random Forest model on test data:')\n",
        "get_pre_rec_f1(rf_best, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4K7WtVx2Sfu7",
        "outputId": "25d91193-89e3-4acd-b270-87aae9cfbfcd"
      },
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Random Forest model on test data:\n",
            "Precision:0.841\n",
            "Recall:0.812\n",
            "F1 score:0.826\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Model Performance:\n",
        "Random Forest has recall rate as 0.812 and F1 score as 0.826. Now we will try another model XGBoost."
      ],
      "metadata": {
        "id": "vbnBnt0_TS0P"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#3. XGBoost"
      ],
      "metadata": {
        "id": "JfCzmrFcTiqd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def run_xgboost(imbalance=False):\n",
        "    # Split data  \n",
        "    X_train, X_test, y_train, y_test = data_split(X, y, imbalance = imbalance)\n",
        "    \n",
        "    # Instantiate model with default parameters\n",
        "    clf_xgb = XGBClassifier()\n",
        "    \n",
        "    # Fit the model\n",
        "    clf_xgb.fit(X_train, y_train)\n",
        "\n",
        "    # Use model's default parameters to get cross validation score\n",
        "    scores = cross_val_score(clf_xgb, X_train, y_train, scoring =\"roc_auc\", cv = 5)\n",
        "    roc_auc_xgb = np.mean(scores)\n",
        "    \n",
        "    if imbalance:\n",
        "        return \"XGBoost\", \"With SMOTE\", roc_auc_xgb\n",
        "    else:\n",
        "        return \"XGBoost\", \"Without SMOTE\", roc_auc_xgb\n",
        "\n",
        "    \n",
        "model_result = []\n",
        "model_result.append(run_xgboost())                  \n",
        "model_result.append(run_xgboost(imbalance=True))    \n",
        "\n",
        "pd.DataFrame(model_result, columns = [\"Model\", \"SMOTE\", \"ROC_AUC Score\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 112
        },
        "id": "nOiqyDTlTetj",
        "outputId": "e039912c-a1fc-4edf-8f0f-368e9dffc067"
      },
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     Model          SMOTE  ROC_AUC Score\n",
              "0  XGBoost  Without SMOTE       0.860241\n",
              "1  XGBoost     With SMOTE       0.860476"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-57d38939-f86a-4f0a-be41-fb6149a62ceb\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>SMOTE</th>\n",
              "      <th>ROC_AUC Score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>XGBoost</td>\n",
              "      <td>Without SMOTE</td>\n",
              "      <td>0.860241</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>XGBoost</td>\n",
              "      <td>With SMOTE</td>\n",
              "      <td>0.860476</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-57d38939-f86a-4f0a-be41-fb6149a62ceb')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-57d38939-f86a-4f0a-be41-fb6149a62ceb button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-57d38939-f86a-4f0a-be41-fb6149a62ceb');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Since SMOTE sampling performed better on training data, we will use SMOTE sampling and initial best guess parameters to fit the model."
      ],
      "metadata": {
        "id": "X4tlN8QBTxz-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Sampling and split data with SMOTE\n",
        "X_train_sm, X_test, y_train_sm, y_test = data_split(X, y, imbalance = True)"
      ],
      "metadata": {
        "id": "KDyqN1zvTzTa"
      },
      "execution_count": 93,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Instantiate models with initial best guess parameters\n",
        "\n",
        "params = { \n",
        "    'gamma':0,\n",
        "    'learning_rate':0.01, \n",
        "    'max_depth':3, \n",
        "    'colsample_bytree':0.6,\n",
        "    'subsample':0.8,\n",
        "    'scale_pos_weight':3.5,\n",
        "    'n_estimators':1000,\n",
        "    'objective':'binary:logistic', \n",
        "    'reg_alpha':0.3    \n",
        "}\n",
        "    \n",
        "# Instantiate model\n",
        "clf_xgb = XGBClassifier(**params)\n",
        "\n",
        "# Use model's initial parameters to get cross validation score\n",
        "scores = cross_val_score(clf_xgb, X_train_sm, y_train_sm, scoring =\"roc_auc\", cv = 5)\n",
        "roc_auc_xgb = np.mean(scores)\n",
        "\n",
        "print(f'ROC_AUC training score with initial best parameters for XGBoost: {roc_auc_xgb:.3f}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j-mNKKCbTKrZ",
        "outputId": "40b17460-074c-46d0-c9e7-82633a8317f3"
      },
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "ROC_AUC training score with initial best parameters for XGBoost: 0.857\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Tune Parameters of XGBoost"
      ],
      "metadata": {
        "id": "24REnQXRUNwH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Number of trees\n",
        "n_estimators = np.arange(200,1000,200)\n",
        "\n",
        "# Minimum loss reduction required to make a further partition on a leaf node of the tree\n",
        "# The larger gamma is, the more conservative the algorithm will be\n",
        "gamma = np.arange(0.1,0.6,0.1)\n",
        "\n",
        "# Default 0.3, range(0,1)\n",
        "learning_rate = np.arange(0.1,0.6,0.1)\n",
        "\n",
        "# Maximum number of levels in tree\n",
        "max_depth = list(range(3,8,1))\n",
        "\n",
        "# Subsample ratio of the training instances.Range(0,1)\n",
        "subsample = np.arange(0.5,0.9,0.1)\n",
        "\n",
        "# Subsample ratio of columns when constructing each tree. Range(0,1)\n",
        "colsample_bytree = np.arange(0.5,0.9,0.1)\n",
        "\n",
        "# Control the balance of positive and negative weights\n",
        "# Sum(negative instances) / sum(positive instances)\n",
        "scale_pos_weight = [1,3.5]\n",
        "\n",
        "\n",
        "# Create the random grid\n",
        "random_grid_xgb = {'n_estimators': n_estimators,\n",
        "                   'gamma': gamma,\n",
        "                   'learning_rate':learning_rate,\n",
        "                   'max_depth': max_depth,\n",
        "                   'subsample':subsample,\n",
        "                   'colsample_bytree':colsample_bytree,\n",
        "                   'scale_pos_weight':scale_pos_weight\n",
        "                  }\n",
        "pprint(random_grid_xgb)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H9sqdnZEUQnO",
        "outputId": "4b3fab5a-6ef4-4a0d-9d11-75fe8da28bb0"
      },
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'colsample_bytree': array([0.5, 0.6, 0.7, 0.8]),\n",
            " 'gamma': array([0.1, 0.2, 0.3, 0.4, 0.5]),\n",
            " 'learning_rate': array([0.1, 0.2, 0.3, 0.4, 0.5]),\n",
            " 'max_depth': [3, 4, 5, 6, 7],\n",
            " 'n_estimators': array([200, 400, 600, 800]),\n",
            " 'scale_pos_weight': [1, 3.5],\n",
            " 'subsample': array([0.5, 0.6, 0.7, 0.8])}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Use randomized search to find best parameters\n",
        "\n",
        "xgboost = XGBClassifier()\n",
        "xgb_random = RandomizedSearchCV(estimator = xgboost, \n",
        "                                param_distributions = random_grid_xgb, \n",
        "                                n_iter = 100, \n",
        "                                cv = 3, \n",
        "                                verbose=1, \n",
        "                                random_state=42, \n",
        "                                n_jobs = -1,\n",
        "                                scoring ='roc_auc')\n",
        "\n",
        "\n",
        "xgb_random.fit(X_train_sm, y_train_sm)   \n",
        "xgb_random.best_params_, xgb_random.best_score_\n",
        "\n",
        "print(xgb_random.best_params_,xgb_random.best_score_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u_u6_1TXUFgw",
        "outputId": "810f3361-0a30-4cd2-deb3-40cfe8df309a"
      },
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n",
            "{'subsample': 0.7999999999999999, 'scale_pos_weight': 1, 'n_estimators': 800, 'max_depth': 7, 'learning_rate': 0.1, 'gamma': 0.4, 'colsample_bytree': 0.7} 0.910821768435926\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save tuned model for future use\n",
        "\n",
        "#joblib.dump(xgb_random,'xgb_random_final.pkl')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zaQhEPJ_U6nv",
        "outputId": "ec6c6bb8-4775-48c4-f0c3-15654b49bd75"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['xgb_random_final.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "After tuning hyperparameters, XGBoost has a ROC_AUC score of 0.857 on training data.\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "##XGBoost Model Evaluation"
      ],
      "metadata": {
        "id": "t4G5Pp3kVhFZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compute precision, recall and F1 score of tuned XGBoost model on test data\n",
        "\n",
        "print('XGBoost model on test data:')\n",
        "\n",
        "get_pre_rec_f1(xgb_random, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VJBAGKAHVd_W",
        "outputId": "67af4633-33f1-434b-d555-37331c06850f"
      },
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "XGBoost model on test data:\n",
            "Precision:0.862\n",
            "Recall:0.832\n",
            "F1 score:0.847\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Model Comparison\n",
        "###Confusion Matrix"
      ],
      "metadata": {
        "id": "TUNDY6X6VvE4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot confusion matrix of 3 models\n",
        "\n",
        "fig,ax=plt.subplots(3,1, figsize=(10,15))\n",
        "\n",
        "plot_confusion_matrix(lr_best, X_test_std, y_test, ax = ax[0], values_format=\"d\")\n",
        "ax[0].set_title(\"Logistic Regression\")\n",
        "\n",
        "plot_confusion_matrix(rf_best, X_test, y_test, ax = ax[1], values_format=\"d\")\n",
        "ax[1].set_title(\"Random Forest\")\n",
        "\n",
        "plot_confusion_matrix(xgb_random, X_test, y_test, ax = ax[2], values_format=\"d\")\n",
        "ax[2].set_title(\"XGBoost\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 910
        },
        "id": "4byFJJVbVudf",
        "outputId": "513112d8-66ea-440d-8368-86c6170b6b40"
      },
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'XGBoost')"
            ]
          },
          "metadata": {},
          "execution_count": 99
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x1080 with 6 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### ROC_AUC Curve"
      ],
      "metadata": {
        "id": "qPw2Z4AiV76n"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot ROC_AUC curve of 3 models\n",
        " \n",
        "fig,ax=plt.subplots(figsize=(10,5))\n",
        "\n",
        "plot_roc_curve(lr_best, X_test_std, y_test,ax=ax, color=\"blue\",label='Logistic Regression')\n",
        "plot_roc_curve(rf_best, X_test, y_test,ax=ax, color=\"black\",label='Random Forest')\n",
        "plot_roc_curve(xgb_random, X_test, y_test,ax=ax, color=\"red\",label='XGBoost')\n",
        "\n",
        "plt.title('ROC/AUC of 3 models')\n",
        "plt.grid()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        },
        "id": "ajak2NlRV4-L",
        "outputId": "cff7e62f-40df-4833-eb4b-b8f43af8bcb8"
      },
      "execution_count": 100,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Precision_Recall Curve\n",
        "\n",
        "\n",
        "---\n",
        "Since the classes are imbalanced, precision_recall curve is more approriate.\n"
      ],
      "metadata": {
        "id": "wfffGtZxWDT4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compare 3 models' Precision_recall curves\n",
        "\n",
        "fig,ax=plt.subplots(figsize=(10,5))\n",
        "\n",
        "plot_precision_recall_curve(lr_best, X_test_std, y_test, ax=ax,color=\"blue\",label='Logistic Regression')\n",
        "plot_precision_recall_curve(rf_best, X_test, y_test, ax=ax,color=\"black\",label='Random Forest')\n",
        "plot_precision_recall_curve(xgb_random, X_test, y_test, ax=ax,color=\"red\",label='XGBoost')\n",
        "\n",
        "plt.title('Precision_Recall of 3 models')\n",
        "plt.grid()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        },
        "id": "uquAhJbgWETc",
        "outputId": "b65c510d-bc07-48bc-e3e5-f4d44c6b1a6d"
      },
      "execution_count": 101,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Compare Models with DummyClassifier"
      ],
      "metadata": {
        "id": "9hVgAAD5k_ZO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.dummy import DummyClassifier\n",
        "\n",
        "dummy_clf = DummyClassifier(strategy=\"stratified\")\n",
        "dummy_clf.fit(X_train, y_train)\n",
        "DummyClassifier(strategy='stratified')\n",
        "y_pred_dummy = dummy_clf.predict(X_test)\n",
        "\n",
        "print('Dummy model:')\n",
        "get_pre_rec_f1(dummy_clf, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iPyxVxwGlBly",
        "outputId": "225a6cc2-af7e-430b-f569-169eee319ed8"
      },
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dummy model:\n",
            "Precision:0.501\n",
            "Recall:0.501\n",
            "F1 score:0.501\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "All 3 models outperformed the dummy classifier's performance. Logistic Regression model has the best recall and Random Forest has the best Precision_Recall balance."
      ],
      "metadata": {
        "id": "W2Cy8clqlLwk"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "| Models|Precision|Recall|F1 Score|Conclusion|\n",
        "| --- | --- | --- | --- |--- |\n",
        "| Dummy Model|0.501|0.501|0.501|Benchmark|\n",
        "| Logistic Regression|0.729|0.714|0.721|Good|\n",
        "| Random Forest|0.841|0.812|0.826|Best model|\n",
        "| XGBoost|0.862|0.832|0.847|Best recall|"
      ],
      "metadata": {
        "id": "-C3yAJonlScG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Model Recommendation\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "We recommend recall = 0.8, however, the threshold can be adjusted to reach higher recall."
      ],
      "metadata": {
        "id": "iTL-iMMqn8Pr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compute precision, recall and threshold of Random Forest\n",
        "\n",
        "y_predict_rf = rf_best.predict_proba(X_test)\n",
        "y_scores_rf = y_predict_rf[:,1]\n",
        "\n",
        "precisions, recalls, thresholds = precision_recall_curve(y_test, y_scores_rf)\n",
        "\n",
        "recalls_80 = recalls[np.argmin(recalls >= 0.80)]               # Recommend recall score = 0.8\n",
        "precision_80 = precisions[np.argmin(recalls >= 0.80)]\n",
        "threshold_80_recall = thresholds[np.argmin(recalls >= 0.80)]\n",
        "\n",
        "thresholds = np.append(thresholds, 1)\n",
        "\n",
        "recalls_80, precision_80, threshold_80_recall"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aCK2bp0KkzL3",
        "outputId": "b821b5bb-ce28-4adf-8386-2d2df1d7ec79"
      },
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.7999714652589528, 0.8509637274244953, 0.5148827833386656)"
            ]
          },
          "metadata": {},
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot recommended recall = 0.8\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10,5))\n",
        "\n",
        "ax.plot(thresholds, recalls, label='Recalls')\n",
        "ax.plot(thresholds, precisions, label='Precisions')\n",
        "\n",
        "ax.plot([threshold_80_recall, threshold_80_recall], [precision_80, recalls_80], \"r:\" )\n",
        "ax.plot([threshold_80_recall, threshold_80_recall], [0, precision_80], \"r:\")\n",
        "ax.plot([0, threshold_80_recall], [precision_80, precision_80], \"r:\")\n",
        "ax.plot([0, threshold_80_recall], [recalls_80, recalls_80], \"r:\")\n",
        "ax.plot([threshold_80_recall], [recalls_80], \"ro\", label='Threshold')\n",
        "\n",
        "ax.set_xlabel('Threshold')\n",
        "ax.set_ylabel('Precisons / Recalls')\n",
        "ax.legend(loc='center right',fontsize=16)\n",
        "plt.title('Model Recommendation(Recalls = 0.8)')\n",
        "plt.savefig('recommended_recall')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        },
        "id": "y14843i8oJ6e",
        "outputId": "d1f50397-5a6f-45a1-bbe8-958b29640580"
      },
      "execution_count": 104,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot feature importance of winner model - Random Forest\n",
        "\n",
        "fea_df = pd.DataFrame({'Feature': feature_cols, 'Feature importance': rf_best.feature_importances_})\n",
        "fea_df = fea_df.sort_values(by='Feature importance')\n",
        "\n",
        "figure, ax = plt.subplots(figsize = (10,8))\n",
        "fea_df.plot.barh(x='Feature',y='Feature importance', ax=ax)\n",
        "plt.title('Features importance',fontsize=14)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "id": "THKnRmQBoQ69",
        "outputId": "109052f5-c3f9-4808-a371-f084718cb27b"
      },
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Features importance')"
            ]
          },
          "metadata": {},
          "execution_count": 105
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\"PAY_1\" AND \"Pay_2\" are the most recent 2 months' payment status and they are the strongest predictors of future payment default risk."
      ],
      "metadata": {
        "id": "tQM60arMoicV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "#Conclusion\n",
        "Logistic Regression model has the highest recall but the lowest precision, if the business cares recall the most, then this model is the best candidate. If the balance of recall and precision is the most important metric, then Random Forest is the ideal model. Since Random Forest has slightly lower recall but much higher precision than Logistic Regression, I would recommend Random Forest."
      ],
      "metadata": {
        "id": "7RU4V7qgol-X"
      }
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.10"
    },
    "colab": {
      "name": "Credit_Card_Default_Prediction.ipynb",
      "provenance": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}