01-intro/homework.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "960293ae-5913-49fe-8a2a-b0c84697e9e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import ptitprince as pt\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "cfeaf7c0-1a01-43a9-ba1e-098702c0f59d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_dataframe(path):\n",
    "    df = pd.read_parquet(path)\n",
    "    df[\"duration_min\"] = (df[\"dropOff_datetime\"] - df[\"pickup_datetime\"]) / (np.timedelta64(1, 's') * 60)\n",
    "    df = df[(df[\"duration_min\"] >= 1) & (df[\"duration_min\"] <= 60)]\n",
    "    \n",
    "    df[\"PUlocationID\"].fillna(-1, inplace=True)\n",
    "    df[\"DOlocationID\"].fillna(-1, inplace=True) \n",
    "    \n",
    "    categorical = ['PUlocationID', 'DOlocationID']\n",
    "    df[categorical] = df[categorical].astype(str)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dae8da3f-fc37-4e5b-a0a5-7e86cf51bac8",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Q1. Downloading the data\n",
    "\n",
    "Read the data for January. How many records are there?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b2df11f0-b4dd-4ccd-9ea5-151ddf30d185",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>dispatching_base_num</th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>dropOff_datetime</th>\n",
       "      <th>PUlocationID</th>\n",
       "      <th>DOlocationID</th>\n",
       "      <th>SR_Flag</th>\n",
       "      <th>Affiliated_base_number</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>B00009</td>\n",
       "      <td>2021-01-01 00:27:00</td>\n",
       "      <td>2021-01-01 00:44:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>B00009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B00009</td>\n",
       "      <td>2021-01-01 00:50:00</td>\n",
       "      <td>2021-01-01 01:07:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>B00009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B00013</td>\n",
       "      <td>2021-01-01 00:01:00</td>\n",
       "      <td>2021-01-01 01:51:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>B00013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B00037</td>\n",
       "      <td>2021-01-01 00:13:09</td>\n",
       "      <td>2021-01-01 00:21:26</td>\n",
       "      <td>NaN</td>\n",
       "      <td>72.0</td>\n",
       "      <td>None</td>\n",
       "      <td>B00037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B00037</td>\n",
       "      <td>2021-01-01 00:38:31</td>\n",
       "      <td>2021-01-01 00:53:44</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61.0</td>\n",
       "      <td>None</td>\n",
       "      <td>B00037</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  dispatching_base_num     pickup_datetime    dropOff_datetime  PUlocationID  \\\n",
       "0               B00009 2021-01-01 00:27:00 2021-01-01 00:44:00           NaN   \n",
       "1               B00009 2021-01-01 00:50:00 2021-01-01 01:07:00           NaN   \n",
       "2               B00013 2021-01-01 00:01:00 2021-01-01 01:51:00           NaN   \n",
       "3               B00037 2021-01-01 00:13:09 2021-01-01 00:21:26           NaN   \n",
       "4               B00037 2021-01-01 00:38:31 2021-01-01 00:53:44           NaN   \n",
       "\n",
       "   DOlocationID SR_Flag Affiliated_base_number  \n",
       "0           NaN    None                 B00009  \n",
       "1           NaN    None                 B00009  \n",
       "2           NaN    None                 B00013  \n",
       "3          72.0    None                 B00037  \n",
       "4          61.0    None                 B00037  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_parquet(\"../data/fhv_tripdata_2021-01.parquet\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "38224164-eb21-4ca2-a4a9-ce702f0d3210",
   "metadata": {},
   "source": [
    "There are 1154112 rows."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84a51fb1-75ee-4cd3-bd3e-4154563c9f85",
   "metadata": {},
   "source": [
    "## Q2. Computing duration\n",
    "\n",
    "Now let's compute the duration variable. It should contain the duration of a ride in minutes.\n",
    "\n",
    "What's the average trip duration in January?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f227f7c9-54c6-4ba6-9659-1a3a057437c5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>dispatching_base_num</th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>dropOff_datetime</th>\n",
       "      <th>PUlocationID</th>\n",
       "      <th>DOlocationID</th>\n",
       "      <th>SR_Flag</th>\n",
       "      <th>Affiliated_base_number</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>B00009</td>\n",
       "      <td>2021-01-01 00:27:00</td>\n",
       "      <td>2021-01-01 00:44:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>B00009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B00009</td>\n",
       "      <td>2021-01-01 00:50:00</td>\n",
       "      <td>2021-01-01 01:07:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>B00009</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  dispatching_base_num     pickup_datetime    dropOff_datetime  PUlocationID  \\\n",
       "0               B00009 2021-01-01 00:27:00 2021-01-01 00:44:00           NaN   \n",
       "1               B00009 2021-01-01 00:50:00 2021-01-01 01:07:00           NaN   \n",
       "\n",
       "   DOlocationID SR_Flag Affiliated_base_number  \n",
       "0           NaN    None                 B00009  \n",
       "1           NaN    None                 B00009  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ad0bfab5-c55c-4c86-bb2c-cc56a313431c",
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"duration_min\"] = (df[\"dropOff_datetime\"] - df[\"pickup_datetime\"]) / (np.timedelta64(1, 's') * 60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "834ad094-e455-4053-b4ff-e1523776f310",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Average trip duration in January is 19.167224093791006'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f\"Average trip duration in January is {df['duration_min'].mean()}\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05f45f12-63d7-49ea-b1b4-3a9da12e331e",
   "metadata": {},
   "source": [
    "## Data preparation\n",
    "Check the distribution of the duration variable. There are some outliners.\n",
    "\n",
    "Let's remove them and keep only the records where the duration was between 1 and 60 minutes (inclusive).\n",
    "\n",
    "How many records did you drop?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "73bf08c1-7638-445a-8c4b-e38ca2a6dd95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='duration_min'>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(x=df[\"duration_min\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c99e1448-e82b-4b14-9505-b515bb4ac4d1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='duration_min'>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.violinplot(x=df[\"duration_min\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d6fed8cd-92d4-4f4b-be18-abaa794c24a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "ort=\"v\"; pal = sns.color_palette(n_colors=1); sigma = 0.8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f078a87a-d864-4b49-8914-3461574e4217",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/mlops-zoomcamp/lib/python3.8/site-packages/seaborn/_core.py:1326: UserWarning: Vertical orientation ignored with only `x` specified.\n",
      "  warnings.warn(single_var_warning.format(\"Vertical\", \"x\"))\n",
      "/opt/anaconda3/envs/mlops-zoomcamp/lib/python3.8/site-packages/seaborn/_core.py:1326: UserWarning: Vertical orientation ignored with only `x` specified.\n",
      "  warnings.warn(single_var_warning.format(\"Vertical\", \"x\"))\n",
      "/opt/anaconda3/envs/mlops-zoomcamp/lib/python3.8/site-packages/seaborn/_core.py:1326: UserWarning: Vertical orientation ignored with only `x` specified.\n",
      "  warnings.warn(single_var_warning.format(\"Vertical\", \"x\"))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(figsize=(7, 5))\n",
    "ax=pt.RainCloud(x=df[\"duration_min\"], palette = pal, bw = sigma,\n",
    "                 width_viol = .9, ax = ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "021650f6-d53e-4388-9fc5-66e5a93c08f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_no_outliers = df[(df[\"duration_min\"] >= 1) & (df[\"duration_min\"] <= 60)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a95481f7-d205-46ff-92f3-1e83cf6a95d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'44286 records were dropped'"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f\"{len(df) - len(df_no_outliers)} records were dropped\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a61190a6-ab1a-4a2f-9268-95d1e6158a8b",
   "metadata": {},
   "source": [
    "## Q3. Missing values\n",
    "\n",
    "The features we'll use for our model are the pickup and dropoff location IDs.\n",
    "\n",
    "But they have a lot of missing values there. Let's replace them with \"-1\"\n",
    "\n",
    "What's the factions of missing values for the pickup location ID? (Or the fraction of \"-1\"s after you filled the NAs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "764a27b8-4b0c-4b7a-afaa-c402e692db4b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['dispatching_base_num', 'pickup_datetime', 'dropOff_datetime',\n",
       "       'PUlocationID', 'DOlocationID', 'SR_Flag', 'Affiliated_base_number',\n",
       "       'duration_min'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "60d0a9a7-35bc-4ce9-855a-2bec5ef7fd60",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(df.isnull(), cbar=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "6c652d24-5c35-43bc-b883-05a1fb9dd843",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Fraction of missing values of pickup location id is: 83.03067639882438'"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f'Fraction of missing values of pickup location id is: {(df[\"PUlocationID\"].isnull().sum() / len(df)) * 100}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "73b98f55-39ac-49f1-af0e-d11c920184e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14.0558282038485"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df[\"DOlocationID\"].isnull().sum() / len(df)) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "a9df0101-0c20-400e-ab0b-17de21fa8c67",
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"PUlocationID\"].fillna(-1, inplace=True) \n",
    "df[\"DOlocationID\"].fillna(-1, inplace=True) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5aa654b-49f0-4ee8-83ea-18bebfa606a8",
   "metadata": {},
   "source": [
    "## Q4. One-hot encoding\n",
    "Let's apply one-hot encoding to the pickup and dropoff location IDs. We'll use only these two features for our model.\n",
    "\n",
    "- Turn the dataframe into a list of dictionaries\n",
    "- Fit a dictionary vectorizer\n",
    "- Get a feature matrix from it\n",
    "\n",
    "What's the dimensionality of this matrix? (The number of columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "fb80c489-be8b-47ec-a42b-4747c81ca84e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.feature_extraction import DictVectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "287805e7-d65f-4426-9ac3-f7b10e3e5015",
   "metadata": {},
   "source": [
    "Convert into categorical variables before encoding."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "fb34df17-25a9-4195-9418-446bf105a4d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "categorical = ['PUlocationID', 'DOlocationID']\n",
    "df[categorical] = df[categorical].astype(str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "af0d1b8a-35da-4d72-aaa9-4f014583446c",
   "metadata": {},
   "outputs": [],
   "source": [
    "v = DictVectorizer(sparse=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "dc11faf1-4e50-4edd-a26b-a3aca3fbf068",
   "metadata": {},
   "outputs": [],
   "source": [
    "list_of_dicts = df[[\"PUlocationID\", \"DOlocationID\"]].to_dict(\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "3f73f7a4-933c-474c-a466-610f993695bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = v.fit_transform(list_of_dicts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "b120b938-713c-4ec5-afa6-c3494d5d91a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1154112, 525)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df29cedd-0ac0-4db7-bd03-a185fa631cd7",
   "metadata": {},
   "source": [
    "Alternative way of doing it with pandas `get_dummies` method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "ca5de478-5423-49b1-a371-bbcad796a70e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_one_hot = pd.get_dummies(df[[\"PUlocationID\", \"DOlocationID\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "e6b5bd55-bde1-449c-b62e-f0fd714fc136",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Numbers of columns in feature matrix is: 525'"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f'Numbers of columns in feature matrix is: {df_one_hot.values.shape[1]}'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c08717f4-f5b3-4d3b-9346-247dfc2e7b71",
   "metadata": {},
   "source": [
    "525 columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82893bf6-3439-4b00-ad10-3ad94ad2cd7a",
   "metadata": {},
   "source": [
    "## Q5. Training a model\n",
    "Now let's use the feature matrix from the previous step to train a model.\n",
    "\n",
    "- Train a plain linear regression model with default parameters\n",
    "- Calculate the RMSE of the model on the training data\n",
    "\n",
    "What's the RMSE on train?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "f11a53cb-899a-4720-a140-3c23dc648b26",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "3985edca-37bc-4955-ba34-b847ef35896d",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = read_dataframe(\"../data/fhv_tripdata_2021-01.parquet\")\n",
    "df_val = read_dataframe(\"../data/fhv_tripdata_2021-02.parquet\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "e9a03c61-ee7f-4ce0-b22b-616bb4741376",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.get_dummies(df_train[[\"PUlocationID\", \"DOlocationID\"]]).values\n",
    "#X_train = dv.fit_transform(train_dicts)\n",
    "\n",
    "X_val = pd.get_dummies(df_val[[\"PUlocationID\", \"DOlocationID\"]]).values\n",
    "#X_val = dv.fit_transform(val_dicts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "0c8e2ffa-6875-4db9-82a8-d27b496b90c6",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "target = \"duration_min\"\n",
    "y_train = df_train[target].values\n",
    "y_val = df_val[target].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "ac48235d-ba8a-4636-8e96-4b443a415aa5",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.528519107168236"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LinearRegression()\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_pred = lr.predict(X_train)\n",
    "\n",
    "mean_squared_error(y_train, y_pred, squared=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c3765b2-3be3-45aa-9b16-a2f389c0ab57",
   "metadata": {},
   "source": [
    "## Q6. Evaluating the model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67035868-7060-45a2-a1c7-84ccad10dcec",
   "metadata": {},
   "source": [
    "Now let's apply this model to the validation dataset (Feb 2021).\n",
    "\n",
    "What's the RMSE on validation?"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eeeebdda-4315-4923-9190-1c66e4126ff4",
   "metadata": {},
   "source": [
    "If I try to do it just as it is, I got an error because feature matrix dimensions are different between training and validation sets."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46fe9ab5-e431-455d-8db2-cb3b7a155401",
   "metadata": {
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "source": [
    "```\n",
    "y_pred = lr.predict(X_val)\n",
    "\n",
    "mean_squared_error(y_val, y_pred, squared=False)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f1e436f3-1807-43c5-8887-00222389f9df",
   "metadata": {},
   "source": [
    "Exploring a little bit, I found that there is one value in `DOlocationID` in validation set that is not present in training set, with 3 observations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "bf92bb64-f0aa-4434-b898-df54b8a2dacc",
   "metadata": {},
   "outputs": [],
   "source": [
    "unique_train = np.concatenate([df_train[\"DOlocationID\"].unique(), df_train[\"PUlocationID\"].unique()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "4c022fec-bece-4d2b-a5ad-76e4ec720e82",
   "metadata": {},
   "outputs": [],
   "source": [
    "unique_val = np.concatenate([df_val[\"DOlocationID\"].unique(), df_val[\"PUlocationID\"].unique()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "7c9c199a-768d-480f-ab4e-2ff0667c742e",
   "metadata": {},
   "outputs": [],
   "source": [
    "set_A = set(df_train[\"DOlocationID\"].unique())\n",
    "set_B = set(df_val[\"DOlocationID\"].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "5444969b-3300-4be5-8f1b-19d58dfd0e68",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "263 264\n"
     ]
    }
   ],
   "source": [
    "print(len(df_train[\"DOlocationID\"].unique()), len(df_val[\"DOlocationID\"].unique()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "5ec8e395-ade3-4478-80bc-f5d88155e74d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'110.0'}"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set_B.difference(set_A)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "c6b3c2e2-de1e-4834-b3c7-0d42be7bc3b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "set_A = set(df_train[\"PUlocationID\"].unique())\n",
    "set_B = set(df_val[\"PUlocationID\"].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "bbdbf5c2-c912-452d-85e9-a1826b8d0376",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "set()"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set_A.difference(set_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "4023833a-2e44-4872-991d-e012e3b1b13f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>dispatching_base_num</th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>dropOff_datetime</th>\n",
       "      <th>PUlocationID</th>\n",
       "      <th>DOlocationID</th>\n",
       "      <th>SR_Flag</th>\n",
       "      <th>Affiliated_base_number</th>\n",
       "      <th>duration_min</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>144951</th>\n",
       "      <td>B03060</td>\n",
       "      <td>2021-02-05 13:18:45</td>\n",
       "      <td>2021-02-05 13:58:48</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>110.0</td>\n",
       "      <td>None</td>\n",
       "      <td>B03060</td>\n",
       "      <td>40.050000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>891678</th>\n",
       "      <td>B00972</td>\n",
       "      <td>2021-02-25 06:51:37</td>\n",
       "      <td>2021-02-25 06:54:21</td>\n",
       "      <td>176.0</td>\n",
       "      <td>110.0</td>\n",
       "      <td>None</td>\n",
       "      <td>B00972</td>\n",
       "      <td>2.733333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>892044</th>\n",
       "      <td>B01452</td>\n",
       "      <td>2021-02-25 06:51:37</td>\n",
       "      <td>2021-02-25 06:54:21</td>\n",
       "      <td>176.0</td>\n",
       "      <td>110.0</td>\n",
       "      <td>None</td>\n",
       "      <td>B01452</td>\n",
       "      <td>2.733333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       dispatching_base_num     pickup_datetime    dropOff_datetime  \\\n",
       "144951               B03060 2021-02-05 13:18:45 2021-02-05 13:58:48   \n",
       "891678               B00972 2021-02-25 06:51:37 2021-02-25 06:54:21   \n",
       "892044               B01452 2021-02-25 06:51:37 2021-02-25 06:54:21   \n",
       "\n",
       "       PUlocationID DOlocationID SR_Flag Affiliated_base_number  duration_min  \n",
       "144951         -1.0        110.0    None                 B03060     40.050000  \n",
       "891678        176.0        110.0    None                 B00972      2.733333  \n",
       "892044        176.0        110.0    None                 B01452      2.733333  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_val[df_val[\"DOlocationID\"] == \"110.0\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1407a6e-4514-496c-887c-8968fb1f1f97",
   "metadata": {},
   "source": [
    "So, to solve this issue, I dropped the value to match the dimensions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2def4915-afb6-48fe-a910-de34a02822f5",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_val_fix = df_val[df_val[\"DOlocationID\"] != \"110.0\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "4c9cad4a-5c09-4685-ba78-155dac7933dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val_fix = pd.get_dummies(df_val_fix[[\"PUlocationID\", \"DOlocationID\"]]).values\n",
    "y_val_fix = df_val_fix[target].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "a2c1110b-2902-46e9-a4a9-6f9f835540f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((990110, 525), (990110,))"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_val_fix.shape, y_val_fix.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "4d107a4d-5b45-4b6a-ac11-d79c999b42a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "11.01425378462221"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = lr.predict(X_val_fix)\n",
    "\n",
    "mean_squared_error(y_val_fix, y_pred, squared=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}