PredictFutureSales.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import datetime\n",
    "from math import sqrt\n",
    "from numpy import loadtxt\n",
    "from itertools import product\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sales_train = pd.read_csv('sales_train.csv')\n",
    "test = pd.read_csv('test.csv')\n",
    "items = pd.read_csv('items.csv')\n",
    "shops = pd.read_csv('shops.csv')\n",
    "item_categories = pd.read_csv('item_categories.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>date</th>\n",
       "      <th>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>59</td>\n",
       "      <td>22154</td>\n",
       "      <td>999.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>03.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>05.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>06.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2554</td>\n",
       "      <td>1709.05</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2555</td>\n",
       "      <td>1099.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  date_block_num  shop_id  item_id  item_price  item_cnt_day\n",
       "0  02.01.2013               0       59    22154      999.00           1.0\n",
       "1  03.01.2013               0       25     2552      899.00           1.0\n",
       "2  05.01.2013               0       25     2552      899.00          -1.0\n",
       "3  06.01.2013               0       25     2554     1709.05           1.0\n",
       "4  15.01.2013               0       25     2555     1099.00           1.0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sales_train.head()\n",
    "# sales_train.item_price.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>item_name</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>! ВО ВЛАСТИ НАВАЖДЕНИЯ (ПЛАСТ.)         D</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>!ABBYY FineReader 12 Professional Edition Full...</td>\n",
       "      <td>1</td>\n",
       "      <td>76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>***В ЛУЧАХ СЛАВЫ   (UNV)                    D</td>\n",
       "      <td>2</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>***ГОЛУБАЯ ВОЛНА  (Univ)                      D</td>\n",
       "      <td>3</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>***КОРОБКА (СТЕКЛО)                       D</td>\n",
       "      <td>4</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           item_name  item_id  \\\n",
       "0          ! ВО ВЛАСТИ НАВАЖДЕНИЯ (ПЛАСТ.)         D        0   \n",
       "1  !ABBYY FineReader 12 Professional Edition Full...        1   \n",
       "2      ***В ЛУЧАХ СЛАВЫ   (UNV)                    D        2   \n",
       "3    ***ГОЛУБАЯ ВОЛНА  (Univ)                      D        3   \n",
       "4        ***КОРОБКА (СТЕКЛО)                       D        4   \n",
       "\n",
       "   item_category_id  \n",
       "0                40  \n",
       "1                76  \n",
       "2                40  \n",
       "3                40  \n",
       "4                40  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "items.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>shop_name</th>\n",
       "      <th>shop_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>!Якутск Орджоникидзе, 56 фран</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>!Якутск ТЦ \"Центральный\" фран</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Адыгея ТЦ \"Мега\"</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Балашиха ТРК \"Октябрь-Киномир\"</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Волжский ТЦ \"Волга Молл\"</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        shop_name  shop_id\n",
       "0   !Якутск Орджоникидзе, 56 фран        0\n",
       "1   !Якутск ТЦ \"Центральный\" фран        1\n",
       "2                Адыгея ТЦ \"Мега\"        2\n",
       "3  Балашиха ТРК \"Октябрь-Киномир\"        3\n",
       "4        Волжский ТЦ \"Волга Молл\"        4"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shops.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge sales_train, items, shios, item_categories\n",
    "df = pd.merge(sales_train, items, on='item_id', how='left')\n",
    "df1 = pd.merge(df, shops, on='shop_id', how='left')\n",
    "df2 = pd.merge(df1, item_categories, on='item_category_id', how='left')\n",
    "df2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop unnecessary columns\n",
    "df_sales = df2.copy()\n",
    "df_sales.drop(['item_id', 'item_category_id', 'shop_id'], axis=1, inplace=True)\n",
    "df_sales.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check duplicates\n",
    "print(df_sales.duplicated().sum())\n",
    "print(df_sales.shape)\n",
    "\n",
    "# drop duplicates\n",
    "df_sales.drop_duplicates(inplace=True)\n",
    "df_sales.duplicated().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check missing values\n",
    "df_sales.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check dataset\n",
    "df_sales.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the abnormal values for item_price and ieem_cnt_day\n",
    "index = df_sales[(df_sales.item_price < 0) | (df_sales.item_cnt_day < 0)].index\n",
    "df_sales.drop(index, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import plotly.express as px\n",
    "import seaborn as sns\n",
    "from plotly.subplots import make_subplots\n",
    "import plotly.graph_objects as go"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a new column for total sales\n",
    "df_sales = df_sales.assign(total_sales = df_sales.item_price * df_sales.item_cnt_day)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_sales.head()\n",
    "df_sales.query('item_cnt_day > 1000')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "monthly_sales = df_sales.pivot_table(index='date_block_num', values='item_cnt_day', aggfunc='sum')\n",
    "monthly_sales.head()\n",
    "\n",
    "#  plot the monthly sales volume\n",
    "# px.scatter(monthly_sales, x=monthly_sales.index, y='item_cnt_day', title='Daily Sales')\n",
    "\n",
    "# #  plot line plot for monthly sales volume\n",
    "# px.line(monthly_total_sales, x=monthly_total_sales.index, y='total_sales', title='monthly_total_sales')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "monthly_revenu = df_sales.pivot_table(index='date_block_num', values='total_sales', aggfunc='sum')\n",
    "monthly_revenu.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compare monthly sales volume and monthly revenue\n",
    "fig, axes = plt.subplots(1, 2, figsize=(16, 8), sharex=False)\n",
    "\n",
    "fig1 = sns.scatterplot(x=monthly_sales.index, y='item_cnt_day', data=monthly_sales, ax=axes[0])\n",
    "fig2 = sns.lineplot(x=monthly_revenu.index, y='total_sales', data=monthly_revenu, ax=axes[1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compare monthly sales volume and monthly revenue\n",
    "fig = make_subplots(rows=1, cols=2, subplot_titles=(\"Monthly Sales Volumn\", \"Monthly Revenu\"))\n",
    "fig.add_trace(go.Scatter(x=monthly_sales.index, y=monthly_sales.item_cnt_day, name='Monthly Sales Volumn'), row=1, col=1)\n",
    "fig.add_trace(go.Line(x=monthly_revenu.index, y=monthly_revenu.total_sales, name='Monthly Revenu'), row=1, col=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that there is a downward trend towards 2015, most likely due to the recession. Meanwhile, we can see that the trend is not linear, and there is a lot of noise in the data. We can also see that there is a seasonal component to the data, with a peak in the summer months.\n",
    "\n",
    "Attention --> The data is not stationary, and we will need to use differencing to make it stationary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  top sales shops\n",
    "sns.countplot(df_sales, y='shop_name', order=df_sales.shop_name.value_counts().iloc[:10].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# show the frequency of sales among shops\n",
    "plt.figure(figsize=(10, 16))\n",
    "sns.countplot(y=df_sales['shop_name'], order=df_sales.shop_name.value_counts().index, palette='Set2')\n",
    "plt.xlabel('number of sold items')\n",
    "plt.ylabel('Shop Name')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.shop_name.value_counts().index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "items_per_shop = df_sales.shop_name.value_counts().to_frame().reset_index().rename(columns={'index': 'shop_name', 'shop_name': 'num_items'})\n",
    "# check the outlier shops\n",
    "px.box(items_per_shop, x='num_items')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the outlier shops\n",
    "index = items_per_shop.query('num_items > 117000').index\n",
    "index "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Number of items sold per shop has a big variance, and we can see that there are some shops that are outliers. We can also see that there are some shops that are selling fewer items than others."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  top sales items\n",
    "sns.countplot(df_sales, y='item_name', order=df_sales.item_name.value_counts().iloc[:10].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# most popular items in each shop \n",
    "# Msoled_items = df_sales.groupby(['shop_name', 'item_name']).size().reset_index(name='count')\n",
    "# Msoled_items\n",
    "\n",
    "# msoled_item_shop.item_name.unique()\n",
    "# msoled_item_shop.item_name.mode().values\n",
    "\n",
    "# get the most sold items in each shop\n",
    "most_sold_items = df_sales.pivot_table(index='shop_name', columns='item_name', values='item_cnt_day', aggfunc='sum')\n",
    "name_item = most_sold_items.idxmax(axis=1)\n",
    "num_item = most_sold_items.max(axis=1)\n",
    "msoled_item_shop = pd.concat([name_item, num_item], axis=1)\n",
    "msoled_item_shop.columns = ['item_name', 'item_cnt_day']\n",
    "msoled_item_shop.sort_values(by='item_cnt_day', ascending=False, inplace=True)\n",
    "msoled_item_shop.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "msoled_item_shop.item_name.mode()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We find that Фирменный пакет майка 1С Интерес белый (34*42) is the most popular item, it might due to the appropriate price and that there are some items that are outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# top sale category\n",
    "sns.countplot(y=df_sales.item_category_name, order=df_sales.item_category_name.value_counts().iloc[:10].index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is interesting to see that the most popular item category is DVD. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# most sold items in each category \n",
    "maxSold_items = df_sales.pivot_table(index='item_category_name', columns='item_name', values='item_cnt_day', aggfunc='sum')\n",
    "# create a new dataframe to store the most sold items in each category\n",
    "name_item = maxSold_items.idxmax(axis=1)\n",
    "num_item = maxSold_items.max(axis=1)\n",
    "df_maxSold_item = pd.concat([name_item, num_item], axis=1)\n",
    "df_maxSold_item.columns = ['item_name', 'item_num']\n",
    "df_maxSold_item.sort_values(by='item_num', ascending=False, inplace=True)\n",
    "df_maxSold_item.head(10)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We would like to know the most popluar item in each category, that would help us to develop a recommandation system regarding up selling and cross selling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rename the columns name of total_sales to revenu\n",
    "df_sales.rename(columns={'total_sales':'revenue'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# most revenue items in each category\n",
    "maxSold_name = df_sales.pivot_table(index='item_category_name', columns='item_name', values='revenue', aggfunc='sum').idxmax(axis=1)\n",
    "maxSold_num = df_sales.pivot_table(index='item_category_name', columns='item_name', values='revenue', aggfunc='sum').max(axis=1)\n",
    "maxSold = pd.concat([maxSold_name, maxSold_num], axis=1)\n",
    "maxSold.columns = ['item_name', 'revenue']\n",
    "maxSold.sort_values(by='revenue', ascending=False, inplace=True)\n",
    "maxSold.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is important to know which item contributes the most to the total revenue for each category. We need to give more attention to these items."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "shop_monthly_sales = sales_train.groupby(['shop_id', 'date_block_num'], as_index=False).agg({'item_cnt_day': 'sum'})\n",
    "shop_monthly_sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "shop_monthly_sales_total = []\n",
    "\n",
    "for shop in shop_monthly_sales.shop_id.unique():\n",
    "    current_shop = shop_monthly_sales.query('shop_id == @shop')\n",
    "    shop_monthly_sales_total.append(go.Scatter(x=current_shop.date_block_num, y=current_shop.item_cnt_day, name=f'shop_id: {shop}'))\n",
    "    \n",
    "layout = go.Layout(title='Monthly Sales of Each Shop', xaxis={'title': 'Month'}, yaxis={'title': 'Sales'})\n",
    "fig = go.Figure(data=shop_monthly_sales_total)\n",
    "iplot(fig)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "En general, there is a downward trend towards 2015. Peaks can be seen around december, we assume that people receive more money thank to the yearly premium and spend more during the new year. Meanwhile, we notice that shop_id 54 is an outlier, it is showing different behavior than the other shops."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.shop_name.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Extract the city name from the shop name and write it to the cities list\n",
    "cities = []\n",
    "for i in df_sales['shop_name']:\n",
    "    city = i.split(' ')[0]\n",
    "    if city[0] == '!':\n",
    "        city = i.split('!')[1].split(' ')[0]\n",
    "        cities.append(city)\n",
    "    elif city == 'Интернет-магазин':\n",
    "        cities.append('Интернет')\n",
    "    elif city == 'Цифровой':\n",
    "        cities.append('Интернет')\n",
    "    else:\n",
    "        cities.append(city)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "city = df_sales.shop_name.unique()[:10]\n",
    "city"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales['city'] = cities\n",
    "df_sales.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "top10City = df_sales.city.value_counts().iloc[:10].to_frame().reset_index()\n",
    "top10City.columns = ['city', 'count']\n",
    "top10City"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot the top 10 cities sales distinguishing by item category\n",
    "\n",
    "px.bar(top10City, x='city', y='count', color='city', title='Top 10 Cities Sales', color_discrete_sequence=px.colors.qualitative.Dark24)\n",
    "# px.bar(top10City, x='city', y='count', title='Top 10 Cities Sales')\n",
    "# sns.barplot(data=df_sales, x='city', y='item_cnt_day')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot the top 10 cities sales distinguishing by item category\n",
    "catSales_per_city = df_sales.groupby(['city', 'item_category_name']).item_cnt_day.sum().to_frame().reset_index().sort_values(by='item_cnt_day', ascending=False).reset_index(drop=True)\n",
    "catSales_per_city"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "top10City_name = top10City.city.unique()\n",
    "catSales_per_city10 = catSales_per_city.query('city in @top10City_name')\n",
    "catSales_per_city10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot the cities sales distinguishing by item category\n",
    "px.bar(catSales_per_city10, x='city', y='item_cnt_day', color='item_category_name', title='Top 10 Cities Sales', color_discrete_sequence=px.colors.qualitative.Dark24)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # get the top 3 item categories in top 10 sales city -- to do \n",
    "\n",
    "# top3cat ={city: df_sales.query('city == @city').item_category_name.value_counts().iloc[:3].to_frame().reset_index().values.tolist() \n",
    "#           for city in df_sales.city.unique()}\n",
    "# top3cat\n",
    "\n",
    "\n",
    "# top3cat_df = pd.DataFrame.from_dict(top3cat, orient='index')\n",
    "# top3cat_df.columns = ['top 1', 'top2', 'top3']\n",
    "# top3cat_df\n",
    "\n",
    "# top3cat = df_sales.item_category_name.value_counts().iloc[:3].to_frame().reset_index()\n",
    "\n",
    "# #  make it to a dataframe\n",
    "# top3cat \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check thhe outlier of the item price\n",
    "sns.boxplot(df_sales.item_price)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the outliers\n",
    "df_sales.drop(df_sales[df_sales.item_price > 250000].index, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.boxplot(df_sales.item_price)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>date</th>\n",
       "      <th>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>59</td>\n",
       "      <td>22154</td>\n",
       "      <td>999.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>03.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>05.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>06.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2554</td>\n",
       "      <td>1709.05</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2555</td>\n",
       "      <td>1099.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  date_block_num  shop_id  item_id  item_price  item_cnt_day\n",
       "0  02.01.2013               0       59    22154      999.00           1.0\n",
       "1  03.01.2013               0       25     2552      899.00           1.0\n",
       "2  05.01.2013               0       25     2552      899.00          -1.0\n",
       "3  06.01.2013               0       25     2554     1709.05           1.0\n",
       "4  15.01.2013               0       25     2555     1099.00           1.0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sales_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2.935849e+06</td>\n",
       "      <td>2.935849e+06</td>\n",
       "      <td>2.935849e+06</td>\n",
       "      <td>2.935849e+06</td>\n",
       "      <td>2.935849e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.456991e+01</td>\n",
       "      <td>3.300173e+01</td>\n",
       "      <td>1.019723e+04</td>\n",
       "      <td>8.908532e+02</td>\n",
       "      <td>1.242641e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>9.422988e+00</td>\n",
       "      <td>1.622697e+01</td>\n",
       "      <td>6.324297e+03</td>\n",
       "      <td>1.729800e+03</td>\n",
       "      <td>2.618834e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-1.000000e+00</td>\n",
       "      <td>-2.200000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7.000000e+00</td>\n",
       "      <td>2.200000e+01</td>\n",
       "      <td>4.476000e+03</td>\n",
       "      <td>2.490000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.400000e+01</td>\n",
       "      <td>3.100000e+01</td>\n",
       "      <td>9.343000e+03</td>\n",
       "      <td>3.990000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.300000e+01</td>\n",
       "      <td>4.700000e+01</td>\n",
       "      <td>1.568400e+04</td>\n",
       "      <td>9.990000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3.300000e+01</td>\n",
       "      <td>5.900000e+01</td>\n",
       "      <td>2.216900e+04</td>\n",
       "      <td>3.079800e+05</td>\n",
       "      <td>2.169000e+03</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       date_block_num       shop_id       item_id    item_price  item_cnt_day\n",
       "count    2.935849e+06  2.935849e+06  2.935849e+06  2.935849e+06  2.935849e+06\n",
       "mean     1.456991e+01  3.300173e+01  1.019723e+04  8.908532e+02  1.242641e+00\n",
       "std      9.422988e+00  1.622697e+01  6.324297e+03  1.729800e+03  2.618834e+00\n",
       "min      0.000000e+00  0.000000e+00  0.000000e+00 -1.000000e+00 -2.200000e+01\n",
       "25%      7.000000e+00  2.200000e+01  4.476000e+03  2.490000e+02  1.000000e+00\n",
       "50%      1.400000e+01  3.100000e+01  9.343000e+03  3.990000e+02  1.000000e+00\n",
       "75%      2.300000e+01  4.700000e+01  1.568400e+04  9.990000e+02  1.000000e+00\n",
       "max      3.300000e+01  5.900000e+01  2.216900e+04  3.079800e+05  2.169000e+03"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sales_train.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>ID</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>5037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>5320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>5233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>5232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>5268</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID  shop_id  item_id\n",
       "0   0        5     5037\n",
       "1   1        5     5320\n",
       "2   2        5     5233\n",
       "3   3        5     5232\n",
       "4   4        5     5268"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. remove the neagtive values\n",
    "2. remove the outliers\n",
    "3. remove the duplicates \n",
    "4. remove unnecessary columns\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# remove the negative values in sales train \n",
    "df_sales = sales_train.copy()\n",
    "index = df_sales[(df_sales.item_cnt_day <0) | (df_sales.item_price <0)].index\n",
    "df_sales.drop(index, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot two subplot of box to check the outliers for item price and item count\n",
    "fig = make_subplots(rows=1, cols=2, subplot_titles=('Item Price', 'Item Count'))\n",
    "fig.add_trace(go.Box(y=df_sales.item_price, name='Item Price'), row=1, col=1)\n",
    "fig.add_trace(go.Box(y=df_sales.item_cnt_day, name='Item Count'), row=1, col=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>ID</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>5037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>5320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>5233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>5232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>5268</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID  shop_id  item_id\n",
       "0   0        5     5037\n",
       "1   1        5     5320\n",
       "2   2        5     5233\n",
       "3   3        5     5232\n",
       "4   4        5     5268"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # select only rows of Novemeber 2013 and Novemeber 2014 to get the result of the test set\n",
    "# train_df = sales_train[sales_train.date_block_num.isin([10, 22])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2935849, 6)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = sales_train.copy()\n",
    "train_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>date</th>\n",
       "      <th>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>59</td>\n",
       "      <td>22154</td>\n",
       "      <td>999.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>03.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>05.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>06.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2554</td>\n",
       "      <td>1709.05</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2555</td>\n",
       "      <td>1099.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  date_block_num  shop_id  item_id  item_price  item_cnt_day\n",
       "0  02.01.2013               0       59    22154      999.00           1.0\n",
       "1  03.01.2013               0       25     2552      899.00           1.0\n",
       "2  05.01.2013               0       25     2552      899.00          -1.0\n",
       "3  06.01.2013               0       25     2554     1709.05           1.0\n",
       "4  15.01.2013               0       25     2555     1099.00           1.0"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# # Eliminate outliers by droping data with standard deviation higher than 3 for item_cnt_day and item_price\n",
    "\n",
    "train_df = train_df[(train_df.item_price < 3 * train_df.item_price.std()) & (train_df.item_cnt_day < 3 * train_df.item_cnt_day.std())]\n",
    "\n",
    "# train_df = train_df[((train_df.item_price - train_df.item_price.mean()) < 3 *train_df.item_price.std()) & ((train_df.item_cnt_day - train_df.item_cnt_day.mean())/ 3* train_df.item_cnt_day.std())]\n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2.440940e+06</td>\n",
       "      <td>2.440940e+06</td>\n",
       "      <td>2.440940e+06</td>\n",
       "      <td>2.440940e+06</td>\n",
       "      <td>2.440940e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.429625e+01</td>\n",
       "      <td>3.311330e+01</td>\n",
       "      <td>1.048877e+04</td>\n",
       "      <td>5.761500e+02</td>\n",
       "      <td>9.949720e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>9.439630e+00</td>\n",
       "      <td>1.628581e+01</td>\n",
       "      <td>6.285502e+03</td>\n",
       "      <td>4.935148e+02</td>\n",
       "      <td>1.023724e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-1.000000e+00</td>\n",
       "      <td>-2.200000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>6.000000e+00</td>\n",
       "      <td>2.200000e+01</td>\n",
       "      <td>4.808000e+03</td>\n",
       "      <td>2.200000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.300000e+01</td>\n",
       "      <td>3.100000e+01</td>\n",
       "      <td>1.019100e+04</td>\n",
       "      <td>3.990000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.200000e+01</td>\n",
       "      <td>4.800000e+01</td>\n",
       "      <td>1.601400e+04</td>\n",
       "      <td>7.990000e+02</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3.300000e+01</td>\n",
       "      <td>5.900000e+01</td>\n",
       "      <td>2.216800e+04</td>\n",
       "      <td>2.354230e+03</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       date_block_num       shop_id       item_id    item_price  item_cnt_day\n",
       "count    2.440940e+06  2.440940e+06  2.440940e+06  2.440940e+06  2.440940e+06\n",
       "mean     1.429625e+01  3.311330e+01  1.048877e+04  5.761500e+02  9.949720e-01\n",
       "std      9.439630e+00  1.628581e+01  6.285502e+03  4.935148e+02  1.023724e-01\n",
       "min      0.000000e+00  0.000000e+00  0.000000e+00 -1.000000e+00 -2.200000e+01\n",
       "25%      6.000000e+00  2.200000e+01  4.808000e+03  2.200000e+02  1.000000e+00\n",
       "50%      1.300000e+01  3.100000e+01  1.019100e+04  3.990000e+02  1.000000e+00\n",
       "75%      2.200000e+01  4.800000e+01  1.601400e+04  7.990000e+02  1.000000e+00\n",
       "max      3.300000e+01  5.900000e+01  2.216800e+04  2.354230e+03  1.000000e+00"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2434865, 6)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# eliminate negative value for item_cnt_day and item_price\n",
    "index = train_df[(train_df.item_price < 0) | (train_df.item_cnt_day < 0)].index\n",
    "train_df.drop(index, inplace=True)\n",
    "train_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "print(train_df.duplicated().sum())\n",
    "# drop the duplicated rows\n",
    "train_df.drop_duplicates(inplace=True)\n",
    "print(train_df.duplicated().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a new column revenue\n",
    "train_df = train_df.assign(revenue = train_df.item_price * train_df.item_cnt_day, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>date</th>\n",
       "      <th>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "      <th>revenue</th>\n",
       "      <th>inplace</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>59</td>\n",
       "      <td>22154</td>\n",
       "      <td>999.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>999.00</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>03.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2552</td>\n",
       "      <td>899.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>899.00</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>06.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2554</td>\n",
       "      <td>1709.05</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1709.05</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2555</td>\n",
       "      <td>1099.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1099.00</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>10.01.2013</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>2564</td>\n",
       "      <td>349.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>349.00</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  date_block_num  shop_id  item_id  item_price  item_cnt_day  \\\n",
       "0  02.01.2013               0       59    22154      999.00           1.0   \n",
       "1  03.01.2013               0       25     2552      899.00           1.0   \n",
       "3  06.01.2013               0       25     2554     1709.05           1.0   \n",
       "4  15.01.2013               0       25     2555     1099.00           1.0   \n",
       "5  10.01.2013               0       25     2564      349.00           1.0   \n",
       "\n",
       "   revenue  inplace  \n",
       "0   999.00     True  \n",
       "1   899.00     True  \n",
       "3  1709.05     True  \n",
       "4  1099.00     True  \n",
       "5   349.00     True  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "date              0\n",
       "date_block_num    0\n",
       "shop_id           0\n",
       "item_id           0\n",
       "item_price        0\n",
       "item_cnt_day      0\n",
       "revenue           0\n",
       "inplace           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train2 = train_df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train2 = df_train2.pivot_table(index=['shop_id', 'item_id'], columns='date_block_num', values='item_cnt_day', aggfunc='sum').fillna(0.0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train2.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>...</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</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",
       "      <td>0.0</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>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>1</th>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</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",
       "      <td>0.0</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>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>2</th>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</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",
       "      <td>0.0</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>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>3</th>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</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",
       "      <td>0.0</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>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>4</th>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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",
       "      <td>0.0</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>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>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "date_block_num  shop_id  item_id    0    1    2    3    4    5    6    7  ...  \\\n",
       "0                     0       31  0.0  5.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "1                     0       32  2.0  4.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "2                     0       33  3.0  3.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "3                     0       35  1.0  8.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "4                     0       36  0.0  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "\n",
       "date_block_num   24   25   26   27   28   29   30   31   32   33  \n",
       "0               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "1               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "2               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "3               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "4               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('int64')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train2.shop_id.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "int64 int64\n"
     ]
    }
   ],
   "source": [
    "# df_train2.shop_id = df_train2.shop_id.astype('str')\n",
    "# df_train2.item_id = df_train2.item_id.astype('str')\n",
    "\n",
    "print(df_train2.shop_id.dtype, df_train2.item_id.dtype)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>...</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</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",
       "      <td>0.0</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>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>1</th>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</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",
       "      <td>0.0</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>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>2</th>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</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",
       "      <td>0.0</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>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>3</th>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</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",
       "      <td>0.0</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>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>4</th>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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",
       "      <td>0.0</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>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>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "date_block_num  shop_id  item_id    0    1    2    3    4    5    6    7  ...  \\\n",
       "0                     0       31  0.0  5.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "1                     0       32  2.0  4.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "2                     0       33  3.0  3.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "3                     0       35  1.0  8.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "4                     0       36  0.0  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   \n",
       "\n",
       "date_block_num   24   25   26   27   28   29   30   31   32   33  \n",
       "0               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "1               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "2               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "3               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "4               0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "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>item_name</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>! ВО ВЛАСТИ НАВАЖДЕНИЯ (ПЛАСТ.)         D</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>!ABBYY FineReader 12 Professional Edition Full...</td>\n",
       "      <td>1</td>\n",
       "      <td>76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>***В ЛУЧАХ СЛАВЫ   (UNV)                    D</td>\n",
       "      <td>2</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>***ГОЛУБАЯ ВОЛНА  (Univ)                      D</td>\n",
       "      <td>3</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>***КОРОБКА (СТЕКЛО)                       D</td>\n",
       "      <td>4</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           item_name  item_id  \\\n",
       "0          ! ВО ВЛАСТИ НАВАЖДЕНИЯ (ПЛАСТ.)         D        0   \n",
       "1  !ABBYY FineReader 12 Professional Edition Full...        1   \n",
       "2      ***В ЛУЧАХ СЛАВЫ   (UNV)                    D        2   \n",
       "3    ***ГОЛУБАЯ ВОЛНА  (Univ)                      D        3   \n",
       "4        ***КОРОБКА (СТЕКЛО)                       D        4   \n",
       "\n",
       "   item_category_id  \n",
       "0                40  \n",
       "1                76  \n",
       "2                40  \n",
       "3                40  \n",
       "4                40  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "items.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>...</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>item_category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</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",
       "      <td>0.0</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>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>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</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",
       "      <td>0.0</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>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>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</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",
       "      <td>0.0</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>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>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</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",
       "      <td>0.0</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>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>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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",
       "      <td>0.0</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>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>37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   shop_id  item_id    0    1    2    3    4    5    6    7  ...   25   26  \\\n",
       "0        0       31  0.0  5.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "1        0       32  2.0  4.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "2        0       33  3.0  3.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "3        0       35  1.0  8.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "4        0       36  0.0  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "\n",
       "    27   28   29   30   31   32   33  item_category_id  \n",
       "0  0.0  0.0  0.0  0.0  0.0  0.0  0.0                37  \n",
       "1  0.0  0.0  0.0  0.0  0.0  0.0  0.0                40  \n",
       "2  0.0  0.0  0.0  0.0  0.0  0.0  0.0                37  \n",
       "3  0.0  0.0  0.0  0.0  0.0  0.0  0.0                40  \n",
       "4  0.0  0.0  0.0  0.0  0.0  0.0  0.0                37  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# add the item_category_id to the df_train\n",
    "df_train2 = pd.merge(df_train2, items[['item_id', 'item_category_id']], on='item_id', how='left')\n",
    "df_train2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "shop_id               int64\n",
       "item_id               int64\n",
       "0                   float64\n",
       "1                   float64\n",
       "2                   float64\n",
       "3                   float64\n",
       "4                   float64\n",
       "5                   float64\n",
       "6                   float64\n",
       "7                   float64\n",
       "8                   float64\n",
       "9                   float64\n",
       "10                  float64\n",
       "11                  float64\n",
       "12                  float64\n",
       "13                  float64\n",
       "14                  float64\n",
       "15                  float64\n",
       "16                  float64\n",
       "17                  float64\n",
       "18                  float64\n",
       "19                  float64\n",
       "20                  float64\n",
       "21                  float64\n",
       "22                  float64\n",
       "23                  float64\n",
       "24                  float64\n",
       "25                  float64\n",
       "26                  float64\n",
       "27                  float64\n",
       "28                  float64\n",
       "29                  float64\n",
       "30                  float64\n",
       "31                  float64\n",
       "32                  float64\n",
       "33                  float64\n",
       "item_category_id      int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train2.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "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>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_category_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>...</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>40</td>\n",
       "      <td>0.0</td>\n",
       "      <td>31.0</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",
       "      <td>...</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",
       "      <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>1</th>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>37</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11.0</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",
       "      <td>...</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",
       "      <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>2</th>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>40</td>\n",
       "      <td>6.0</td>\n",
       "      <td>10.0</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",
       "      <td>...</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",
       "      <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>3</th>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "      <td>37</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</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",
       "      <td>...</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",
       "      <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>4</th>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>40</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.0</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",
       "      <td>...</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",
       "      <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>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   shop_id  item_id  item_category_id    0     1    2    3    4    5    6  \\\n",
       "0        0       30                40  0.0  31.0  0.0  0.0  0.0  0.0  0.0   \n",
       "1        0       31                37  0.0  11.0  0.0  0.0  0.0  0.0  0.0   \n",
       "2        0       32                40  6.0  10.0  0.0  0.0  0.0  0.0  0.0   \n",
       "3        0       33                37  3.0   3.0  0.0  0.0  0.0  0.0  0.0   \n",
       "4        0       35                40  1.0  14.0  0.0  0.0  0.0  0.0  0.0   \n",
       "\n",
       "   ...   24   25   26   27   28   29   30   31   32   33  \n",
       "0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "1  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "2  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "3  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "4  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Rearrange the columns\n",
    "df_train2 = df_train2[['shop_id', 'item_id', 'item_category_id'] + list(range(34))]\n",
    "df_train2.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 1 for XGBregressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost import plot_tree\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "from xgboost import XGBRegressor, plot_importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the data into train and test\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_train2.drop([33], axis=1), df_train2[33], test_size=0.2, random_state=42)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_model = XGBRegressor(\n",
    "    max_depth=8,\n",
    "    n_estimators=800,\n",
    "    min_child_weight=300, \n",
    "    colsample_bytree=0.8, \n",
    "    subsample=0.8, \n",
    "    eta=0.3,    \n",
    "    seed=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python310\\lib\\site-packages\\xgboost\\sklearn.py:861: UserWarning: `early_stopping_rounds` in `fit` method is deprecated for better compatibility with scikit-learn, use `early_stopping_rounds` in constructor or`set_params` instead.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training: Done!\n",
      "Predictions: Done!\n"
     ]
    }
   ],
   "source": [
    "xgb_model.fit(\n",
    "    X_train, y_train\n",
    "    , early_stopping_rounds=5\n",
    "    , eval_set=[(X_test, y_test)]\n",
    "    , verbose=False\n",
    ")\n",
    "print('Training: Done!')\n",
    "\n",
    "xgb_predictions = xgb_model.predict(X_test)\n",
    "print('Predictions: Done!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot: title={'center': 'Feature importance'}, xlabel='F score', ylabel='Features'>"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x2000 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Take a look at the importance of the features in the dataset\n",
    "fig, ax = plt.subplots(figsize = (20, 20))\n",
    "plot_importance(xgb_model, ax=ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 0.3824910281897971\n"
     ]
    }
   ],
   "source": [
    "# Evaluating\n",
    "xgb_rmse = sqrt(mean_squared_error(xgb_predictions, y_test))\n",
    "print('RMSE:', xgb_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = pd.read_csv('test.csv')\n",
    "test_df = test.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(214200, 3)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = test_df.merge(df_train2, on=['shop_id', 'item_id'], how='left').fillna(0.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df.drop('ID', axis=1, inplace=True)\n",
    "test_df.drop(33, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "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>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>...</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>item_category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5</td>\n",
       "      <td>5037</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",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5</td>\n",
       "      <td>5320</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",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>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>2</th>\n",
       "      <td>5</td>\n",
       "      <td>5233</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",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>5232</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",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>23.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>5268</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",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>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>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   shop_id  item_id    0    1    2    3    4    5    6    7  ...   24   25  \\\n",
       "0        5     5037  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  2.0  0.0   \n",
       "1        5     5320  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "2        5     5233  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "3        5     5232  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "4        5     5268  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0   \n",
       "\n",
       "    26   27   28   29   30   31   32  item_category_id  \n",
       "0  0.0  0.0  1.0  1.0  1.0  3.0  1.0              19.0  \n",
       "1  0.0  0.0  0.0  0.0  0.0  0.0  0.0               0.0  \n",
       "2  0.0  0.0  1.0  2.0  0.0  1.0  3.0              19.0  \n",
       "3  0.0  0.0  0.0  0.0  0.0  1.0  0.0              23.0  \n",
       "4  0.0  0.0  0.0  0.0  0.0  0.0  0.0               0.0  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(214200, 36)"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predict the test set\n",
    "preds = xgb_model.predict(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "output = pd.DataFrame({\n",
    "    'ID': test.index\n",
    "    ,  'item_cnt_month': preds\n",
    "})\n",
    "\n",
    "output.to_csv('submission3.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 2 for RandomForestRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "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>date</th>\n",
       "      <th>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_price</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1027580</th>\n",
       "      <td>06.11.2013</td>\n",
       "      <td>10</td>\n",
       "      <td>43</td>\n",
       "      <td>4420</td>\n",
       "      <td>299.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1027581</th>\n",
       "      <td>16.11.2013</td>\n",
       "      <td>10</td>\n",
       "      <td>28</td>\n",
       "      <td>10033</td>\n",
       "      <td>199.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1027582</th>\n",
       "      <td>14.11.2013</td>\n",
       "      <td>10</td>\n",
       "      <td>28</td>\n",
       "      <td>10051</td>\n",
       "      <td>149.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1027583</th>\n",
       "      <td>22.11.2013</td>\n",
       "      <td>10</td>\n",
       "      <td>28</td>\n",
       "      <td>10051</td>\n",
       "      <td>149.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1027584</th>\n",
       "      <td>30.11.2013</td>\n",
       "      <td>10</td>\n",
       "      <td>28</td>\n",
       "      <td>10051</td>\n",
       "      <td>149.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               date  date_block_num  shop_id  item_id  item_price  \\\n",
       "1027580  06.11.2013              10       43     4420       299.0   \n",
       "1027581  16.11.2013              10       28    10033       199.0   \n",
       "1027582  14.11.2013              10       28    10051       149.0   \n",
       "1027583  22.11.2013              10       28    10051       149.0   \n",
       "1027584  30.11.2013              10       28    10051       149.0   \n",
       "\n",
       "         item_cnt_day  \n",
       "1027580           1.0  \n",
       "1027581           1.0  \n",
       "1027582           1.0  \n",
       "1027583           1.0  \n",
       "1027584           1.0  "
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## select row of Novemeber 2013 and Novemeber 2014 to get the result of the test set\n",
    "train_df = sales_train[sales_train.date_block_num.isin([10, 22])]\n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>33</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>97</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>482</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>791</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>944</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   date_block_num  shop_id  item_id  item_cnt_day\n",
       "0              10        2       33           2.0\n",
       "1              10        2       97           1.0\n",
       "2              10        2      482           2.0\n",
       "3              10        2      791           1.0\n",
       "4              10        2      944           1.0"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.groupby(['date_block_num', 'shop_id', 'item_id'], as_index=False).item_cnt_day.sum()\n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = train_df.copy()\n",
    "y = train_df.iloc[:, :-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_cnt_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>33</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>97</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>482</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>791</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>944</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   date_block_num  shop_id  item_id  item_cnt_day\n",
       "0              10        2       33           2.0\n",
       "1              10        2       97           1.0\n",
       "2              10        2      482           2.0\n",
       "3              10        2      791           1.0\n",
       "4              10        2      944           1.0"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         2.0\n",
       "1         1.0\n",
       "2         2.0\n",
       "3         1.0\n",
       "4         1.0\n",
       "         ... \n",
       "97210     1.0\n",
       "97211     6.0\n",
       "97212    15.0\n",
       "97213     1.0\n",
       "97214     3.0\n",
       "Name: item_cnt_day, Length: 97215, dtype: float64"
      ]
     },
     "execution_count": 194,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = X.pop('item_cnt_day')\n",
    "y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 2 for RandomForestRegressor with other features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.metrics import mean_squared_error \n",
    "from sklearn.ensemble import RandomForestRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0\n",
    "                                                    , train_size=0.80\n",
    "                                                    , test_size=0.20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating Numerical Transformer\n",
    "numerical_transformer = Pipeline(steps=[\n",
    "    ('scaler', StandardScaler())\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Bundling Preprocessors\n",
    "preprocessor = ColumnTransformer(\n",
    "    transformers=[\n",
    "        ('numerical', numerical_transformer, X.columns),\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating Model\n",
    "rfg_model = RandomForestRegressor(\n",
    "    n_estimators=200,\n",
    "    max_depth=10,\n",
    "    min_samples_leaf=5,\n",
    "    criterion='squared_error',\n",
    "    random_state=0,\n",
    ")\n",
    "\n",
    "rfg_pipeline = Pipeline(steps=[\n",
    "    ('preprocessor', preprocessor),\n",
    "    ('model', rfg_model),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training: Done!\n",
      "Predictions: Done!\n"
     ]
    }
   ],
   "source": [
    "# Training and Predicting\n",
    "rfg_pipeline.fit(X_train, y_train)\n",
    "print('Training: Done!')\n",
    "rfg_predictions = rfg_pipeline.predict(X_test)\n",
    "print('Predictions: Done!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RFG RMSE: 7.061070361673437\n"
     ]
    }
   ],
   "source": [
    "# Calculating RMSE\n",
    "from math import sqrt\n",
    "rfg_rmse = np.sqrt(mean_squared_error(rfg_predictions, y_test))\n",
    "print('RFG RMSE:', rfg_rmse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 3 for XGBregressor with other features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating the Model\n",
    "xgb_model = XGBRegressor(\n",
    "    n_estimators=300,\n",
    "    learning_rate=0.02,\n",
    "    n_jobs=4,\n",
    "    max_depth=8,\n",
    "    min_child_weight=300\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "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>date_block_num</th>\n",
       "      <th>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>85020</th>\n",
       "      <td>22</td>\n",
       "      <td>47</td>\n",
       "      <td>14419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90976</th>\n",
       "      <td>22</td>\n",
       "      <td>54</td>\n",
       "      <td>13923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9830</th>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "      <td>10904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82724</th>\n",
       "      <td>22</td>\n",
       "      <td>44</td>\n",
       "      <td>21909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73660</th>\n",
       "      <td>22</td>\n",
       "      <td>31</td>\n",
       "      <td>11217</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21243</th>\n",
       "      <td>10</td>\n",
       "      <td>28</td>\n",
       "      <td>5323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45891</th>\n",
       "      <td>10</td>\n",
       "      <td>56</td>\n",
       "      <td>2831</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42613</th>\n",
       "      <td>10</td>\n",
       "      <td>54</td>\n",
       "      <td>1600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43567</th>\n",
       "      <td>10</td>\n",
       "      <td>54</td>\n",
       "      <td>9383</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68268</th>\n",
       "      <td>22</td>\n",
       "      <td>27</td>\n",
       "      <td>14117</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>77772 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       date_block_num  shop_id  item_id\n",
       "85020              22       47    14419\n",
       "90976              22       54    13923\n",
       "9830               10       17    10904\n",
       "82724              22       44    21909\n",
       "73660              22       31    11217\n",
       "...               ...      ...      ...\n",
       "21243              10       28     5323\n",
       "45891              10       56     2831\n",
       "42613              10       54     1600\n",
       "43567              10       54     9383\n",
       "68268              22       27    14117\n",
       "\n",
       "[77772 rows x 3 columns]"
      ]
     },
     "execution_count": 233,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python310\\lib\\site-packages\\xgboost\\sklearn.py:861: UserWarning:\n",
      "\n",
      "`early_stopping_rounds` in `fit` method is deprecated for better compatibility with scikit-learn, use `early_stopping_rounds` in constructor or`set_params` instead.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training: Done!\n",
      "Predictions: Done!\n"
     ]
    }
   ],
   "source": [
    "# Traning the Model and Making Predictions\n",
    "\n",
    "xgb_model.fit(\n",
    "    X_train, y_train\n",
    "    , early_stopping_rounds=5\n",
    "    , eval_set=[(X_test, y_test)]\n",
    "    , verbose=False\n",
    ")\n",
    "print('Training: Done!')\n",
    "\n",
    "xgb_predictions = xgb_model.predict(X_test)\n",
    "print('Predictions: Done!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 8.620001609768979\n"
     ]
    }
   ],
   "source": [
    "# Evaluating\n",
    "xgb_rmse = sqrt(mean_squared_error(xgb_predictions, y_test))\n",
    "print('RMSE:', xgb_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "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>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5</td>\n",
       "      <td>5037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5</td>\n",
       "      <td>5320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>5233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>5232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>5268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214195</th>\n",
       "      <td>45</td>\n",
       "      <td>18454</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214196</th>\n",
       "      <td>45</td>\n",
       "      <td>16188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214197</th>\n",
       "      <td>45</td>\n",
       "      <td>15757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214198</th>\n",
       "      <td>45</td>\n",
       "      <td>19648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214199</th>\n",
       "      <td>45</td>\n",
       "      <td>969</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>214200 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        shop_id  item_id\n",
       "0             5     5037\n",
       "1             5     5320\n",
       "2             5     5233\n",
       "3             5     5232\n",
       "4             5     5268\n",
       "...         ...      ...\n",
       "214195       45    18454\n",
       "214196       45    16188\n",
       "214197       45    15757\n",
       "214198       45    19648\n",
       "214199       45      969\n",
       "\n",
       "[214200 rows x 2 columns]"
      ]
     },
     "execution_count": 236,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating Clusters in the Test Dataset using K-Means\n",
    "test.pop('ID')\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>shop_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>date_block_num</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5</td>\n",
       "      <td>5037</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5</td>\n",
       "      <td>5320</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>5233</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>5232</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>5268</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shop_id  item_id  date_block_num\n",
       "0        5     5037              34\n",
       "1        5     5320              34\n",
       "2        5     5233              34\n",
       "3        5     5232              34\n",
       "4        5     5268              34"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Adding November 2015 (34th month)\n",
    "test['date_block_num'] = 34\n",
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Making Predictions and Saving Them into a File\n",
    "predictions_test = rfg_pipeline.predict(test)\n",
    "output = pd.DataFrame({\n",
    "    'ID': test.index\n",
    "    ,  'item_cnt_month': predictions_test\n",
    "})\n",
    "\n",
    "output.to_csv('submission.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  try KMeans Model \n",
    "# from sklearn.cluster import KMeans\n",
    "# criteria = []\n",
    "\n",
    "# for i in range(1, 11):\n",
    "#     kmeans = KMeans(n_clusters=i, init='k-means++', random_state=47, n_init=10)\n",
    "#     kmeans.fit(X)\n",
    "#     criteria.append(kmeans.inertia_)\n",
    "    \n",
    "    \n",
    "# # plotting de criteria\n",
    "# plt.plot(range(1, 11), criteria)\n",
    "# plt.title('The Elbow Method')\n",
    "# plt.xlabel('Number of clusters')\n",
    "# plt.ylabel('Criteria')\n",
    "# plt.show()\n",
    "\n",
    "\n",
    "\n",
    "# # We can see that the elbow is at 4, so we will use 4 clusters\n",
    "\n",
    "# kmeans = KMeans(n_clusters=4, init='k-means++', random_state=47, n_init=10)\n",
    "\n",
    "# X_temps = X.copy()\n",
    "# X_temps['cluster'] = kmeans.fit_predict(X)\n",
    "# X_temps\n",
    "\n",
    "\n",
    "# sns.relplot(x='item_id', y='shop_id', hue='cluster', data=X_temps, height=6)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.7 64-bit",
   "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": "3.10.7"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "369f2c481f4da34e4445cda3fffd2e751bd1c4d706f27375911949ba6bb62e1c"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}