Machine Learning for Retail Sales Forecasting — Features Engineering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Notebook inspired by the models shared:\n",
    "# Link: https://www.kaggle.com/anshuls235/time-series-forecasting-eda-fe-modelling\n",
    "# Link: https://www.kaggle.com/kneroma/m5-first-public-notebook-under-0-50/notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\dask\\dataframe\\utils.py:15: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
      "  import pandas.util.testing as tm\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import lightgbm\n",
    "from lightgbm import LGBMRegressor\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "import joblib\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "import pickle\n",
    "from pathlib import Path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Better rendering \n",
    "from IPython.core.display import HTML\n",
    "HTML(\"<style>.rendered_html th {max-width: 120px;}</style>\")\n",
    "\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "# settings to display all columns\n",
    "pd.set_option(\"display.max_columns\", None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>\n",
       ".dataframe td {\n",
       "    white-space: nowrap;\n",
       "}\n",
       "</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%html\n",
    "<style>\n",
    ".dataframe td {\n",
    "    white-space: nowrap;\n",
    "}\n",
    "</style>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Fetch Data and Processing\n",
    "#### Import Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30,490 records for training data\n",
      "1,969 records for calendar data\n",
      "6,841,121 records for price data\n"
     ]
    }
   ],
   "source": [
    "# Sales Data: Training Set (1-1913) + Validation Set (1914-1941)\n",
    "sales = pd.read_csv('Data/sales_train_evaluation.csv')\n",
    "sales.name = 'sales'\n",
    "\n",
    "print(\"{:,} records for training data\".format(len(sales)))\n",
    "\n",
    "# Calendar\n",
    "calendar = pd.read_csv('Data/calendar.csv')\n",
    "calendar.name = 'calendar'\n",
    "print(\"{:,} records for calendar data\".format(len(calendar)))\n",
    "\n",
    "# Pricing\n",
    "prices = pd.read_csv('Data/sells_prices_improved.csv')\n",
    "prices.name = 'prices'\n",
    "print(\"{:,} records for price data\".format(len(prices)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add records for testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Add zero sales for the remaining days 1942-1969\n",
    "for d in range(1942,1970):\n",
    "    col = 'd_' + str(d)\n",
    "    sales[col] = 0\n",
    "    sales[col] = sales[col].astype(np.int16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Downcasting to reduce memory\n",
    "Source: https://www.kaggle.com/anshuls235/time-series-forecasting-eda-fe-modelling\n",
    "> Shout out to him, it's very useful!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Downcast in order to save memory\n",
    "def downcast(df):\n",
    "    cols = df.dtypes.index.tolist()\n",
    "    types = df.dtypes.values.tolist()\n",
    "    for i,t in enumerate(types):\n",
    "        # Integer\n",
    "        if 'int' in str(t):\n",
    "            # Check if minimum and maximum are in the limit of int8\n",
    "            if df[cols[i]].min() > np.iinfo(np.int8).min and df[cols[i]].max() < np.iinfo(np.int8).max:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.int8)\n",
    "            # Check if minimum and maximum are in the limit of int16\n",
    "            elif df[cols[i]].min() > np.iinfo(np.int16).min and df[cols[i]].max() < np.iinfo(np.int16).max:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.int16)\n",
    "            # Check if minimum and maximum are in the limit of int32\n",
    "            elif df[cols[i]].min() > np.iinfo(np.int32).min and df[cols[i]].max() < np.iinfo(np.int32).max:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.int32)\n",
    "            # Choose int64\n",
    "            else:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.int64)\n",
    "        # Float\n",
    "        elif 'float' in str(t):\n",
    "            if df[cols[i]].min() > np.finfo(np.float16).min and df[cols[i]].max() < np.finfo(np.float16).max:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.float16)\n",
    "            elif df[cols[i]].min() > np.finfo(np.float32).min and df[cols[i]].max() < np.finfo(np.float32).max:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.float32)\n",
    "            else:\n",
    "                df[cols[i]] = df[cols[i]].astype(np.float64)\n",
    "        # Object\n",
    "        elif t == np.object:\n",
    "            if cols[i] == 'date':\n",
    "                df[cols[i]] = pd.to_datetime(df[cols[i]], format='%Y-%m-%d')\n",
    "            else:\n",
    "                df[cols[i]] = df[cols[i]].astype('category')\n",
    "    return df  \n",
    "\n",
    "# # Apply downcasting\n",
    "sales = downcast(sales)\n",
    "prices = downcast(prices)\n",
    "calendar = downcast(calendar)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Additional Features: Pricing\n",
    "- Relative difference between the current price of an item and its historical average price, to highlight promotional offers’ impact.\n",
    "- Price relative difference with the same item sold in other stores, to understand whether or not the store has an attractive price.\n",
    "- Price relative difference with other items sold in the same store and same product category, to capture some cannibalization effects."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def improve_price():\n",
    "    # Calculate Average price for all stores\n",
    "    df_mean_store = pd.DataFrame(prices.groupby(['item_id', 'wm_yr_wk'])['sell_price'].mean())\n",
    "    df_mean_store.columns = ['item_sells_price_avg']\n",
    "    df_mean_store.reset_index(inplace = True)\n",
    "\n",
    "    # Combine with calendar\n",
    "    prices_new = pd.merge(prices, df_mean_store, on=['item_id', 'wm_yr_wk'], how='left', suffixes=('', '_y'))\n",
    "    prices_new.drop(prices_new.filter(regex='_y$').columns.tolist(),axis=1, inplace=True)\n",
    "\n",
    "    # Price difference with same items in other stores\n",
    "    prices_new['delta_price_all_rel'] = (prices_new['sell_price'] - \n",
    "                                         prices_new['item_sells_price_avg'])/prices_new['item_sells_price_avg']\n",
    "\n",
    "    # Price difference with last week\n",
    "    prices_new['item_store'] = prices_new['item_id'].astype(str) + '_' + prices_new['store_id'].astype(str)\n",
    "    prices_new['item_store_change'] = prices_new[\"item_store\"].shift() != prices_new[\"item_store\"]\n",
    "    # Price difference week n - week n-1\n",
    "    prices_new['delta_price_weekn-1'] = (prices_new['sell_price']-\n",
    "                                         prices_new['sell_price'].shift(1)).fillna(0)/prices_new['sell_price'].shift(1)\n",
    "    prices_new['delta_price_weekn-1'] = prices_new['delta_price_weekn-1'].fillna(0) * (prices_new['item_store_change']==0)\n",
    "\n",
    "    # Average price of the department by store\n",
    "    prices_new['dept_id'] = prices_new.item_id.str[:-4]\n",
    "    df_mean_cat = pd.DataFrame(prices_new.groupby(['dept_id', 'store_id', 'wm_yr_wk'])['sell_price'].mean())\n",
    "    df_mean_cat.columns = ['dept_sells_price_avg']\n",
    "    df_mean_cat.reset_index(inplace = True)\n",
    "    # Combine with price dataset\n",
    "    prices_new = pd.merge(prices_new, df_mean_cat, on=['dept_id', 'store_id', 'wm_yr_wk']\n",
    "                          , how='left', suffixes=('', '_y'))\n",
    "    prices_new.drop(prices_new.filter(regex='_y$').columns.tolist(),axis=1, inplace=True)\n",
    "\n",
    "    # Cannibalisation: compare this item price with average of the department (category)\n",
    "    prices_new['delta_price_cat_rel'] = (prices_new['sell_price'] - \n",
    "                                         prices_new['dept_sells_price_avg'])/prices_new['dept_sells_price_avg']                                               \n",
    "\n",
    "    # Drop columns\n",
    "    prices_new.drop(['item_sells_price_avg', 'item_store_change', 'item_store_change', 'dept_id', 'item_store',\n",
    "                    'dept_sells_price_avg'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Melt Data to reconstitute sales records\n",
    "#### Convert from wide to long format (records on row level)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60,034,810 records to combine with calendar and price\n"
     ]
    },
    {
     "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>item_id</th>\n",
       "      <th>dept_id</th>\n",
       "      <th>cat_id</th>\n",
       "      <th>store_id</th>\n",
       "      <th>state_id</th>\n",
       "      <th>d</th>\n",
       "      <th>sold</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HOBBIES_1_001_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_001</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HOBBIES_1_002_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_002</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HOBBIES_1_003_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_003</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HOBBIES_1_004_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_004</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HOBBIES_1_005_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_005</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              id        item_id    dept_id   cat_id store_id  \\\n",
       "0  HOBBIES_1_001_CA_1_evaluation  HOBBIES_1_001  HOBBIES_1  HOBBIES     CA_1   \n",
       "1  HOBBIES_1_002_CA_1_evaluation  HOBBIES_1_002  HOBBIES_1  HOBBIES     CA_1   \n",
       "2  HOBBIES_1_003_CA_1_evaluation  HOBBIES_1_003  HOBBIES_1  HOBBIES     CA_1   \n",
       "3  HOBBIES_1_004_CA_1_evaluation  HOBBIES_1_004  HOBBIES_1  HOBBIES     CA_1   \n",
       "4  HOBBIES_1_005_CA_1_evaluation  HOBBIES_1_005  HOBBIES_1  HOBBIES     CA_1   \n",
       "\n",
       "  state_id    d  sold  \n",
       "0       CA  d_1     0  \n",
       "1       CA  d_1     0  \n",
       "2       CA  d_1     0  \n",
       "3       CA  d_1     0  \n",
       "4       CA  d_1     0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melt\n",
    "df = pd.melt(sales, id_vars=['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'], \n",
    "             var_name='d', value_name='sold').dropna()\n",
    "print(\"{:,} records to combine with calendar and price\".format(len(df)))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combine Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Combine with calendar\n",
    "df = pd.merge(df, calendar, on='d', how='left')\n",
    "# Combine with price\n",
    "df = pd.merge(df, prices, on=['store_id','item_id','wm_yr_wk'], how='left') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60,034,810 records in df\n"
     ]
    },
    {
     "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>item_id</th>\n",
       "      <th>dept_id</th>\n",
       "      <th>cat_id</th>\n",
       "      <th>store_id</th>\n",
       "      <th>state_id</th>\n",
       "      <th>d</th>\n",
       "      <th>sold</th>\n",
       "      <th>date</th>\n",
       "      <th>wm_yr_wk</th>\n",
       "      <th>weekday</th>\n",
       "      <th>wday</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>event_name_1</th>\n",
       "      <th>event_type_1</th>\n",
       "      <th>event_name_2</th>\n",
       "      <th>event_type_2</th>\n",
       "      <th>snap_CA</th>\n",
       "      <th>snap_TX</th>\n",
       "      <th>snap_WI</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>sell_price</th>\n",
       "      <th>delta_price_all_rel</th>\n",
       "      <th>delta_price_weekn-1</th>\n",
       "      <th>delta_price_cat_rel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HOBBIES_1_001_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_001</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "      <td>2011-01-29</td>\n",
       "      <td>11101</td>\n",
       "      <td>Saturday</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HOBBIES_1_002_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_002</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "      <td>2011-01-29</td>\n",
       "      <td>11101</td>\n",
       "      <td>Saturday</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HOBBIES_1_003_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_003</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "      <td>2011-01-29</td>\n",
       "      <td>11101</td>\n",
       "      <td>Saturday</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HOBBIES_1_004_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_004</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "      <td>2011-01-29</td>\n",
       "      <td>11101</td>\n",
       "      <td>Saturday</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HOBBIES_1_005_CA_1_evaluation</td>\n",
       "      <td>HOBBIES_1_005</td>\n",
       "      <td>HOBBIES_1</td>\n",
       "      <td>HOBBIES</td>\n",
       "      <td>CA_1</td>\n",
       "      <td>CA</td>\n",
       "      <td>d_1</td>\n",
       "      <td>0</td>\n",
       "      <td>2011-01-29</td>\n",
       "      <td>11101</td>\n",
       "      <td>Saturday</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              id        item_id    dept_id   cat_id store_id  \\\n",
       "0  HOBBIES_1_001_CA_1_evaluation  HOBBIES_1_001  HOBBIES_1  HOBBIES     CA_1   \n",
       "1  HOBBIES_1_002_CA_1_evaluation  HOBBIES_1_002  HOBBIES_1  HOBBIES     CA_1   \n",
       "2  HOBBIES_1_003_CA_1_evaluation  HOBBIES_1_003  HOBBIES_1  HOBBIES     CA_1   \n",
       "3  HOBBIES_1_004_CA_1_evaluation  HOBBIES_1_004  HOBBIES_1  HOBBIES     CA_1   \n",
       "4  HOBBIES_1_005_CA_1_evaluation  HOBBIES_1_005  HOBBIES_1  HOBBIES     CA_1   \n",
       "\n",
       "  state_id    d  sold       date  wm_yr_wk   weekday  wday  month  year  \\\n",
       "0       CA  d_1     0 2011-01-29     11101  Saturday     1      1  2011   \n",
       "1       CA  d_1     0 2011-01-29     11101  Saturday     1      1  2011   \n",
       "2       CA  d_1     0 2011-01-29     11101  Saturday     1      1  2011   \n",
       "3       CA  d_1     0 2011-01-29     11101  Saturday     1      1  2011   \n",
       "4       CA  d_1     0 2011-01-29     11101  Saturday     1      1  2011   \n",
       "\n",
       "  event_name_1 event_type_1 event_name_2 event_type_2  snap_CA  snap_TX  \\\n",
       "0          NaN          NaN          NaN          NaN        0        0   \n",
       "1          NaN          NaN          NaN          NaN        0        0   \n",
       "2          NaN          NaN          NaN          NaN        0        0   \n",
       "3          NaN          NaN          NaN          NaN        0        0   \n",
       "4          NaN          NaN          NaN          NaN        0        0   \n",
       "\n",
       "   snap_WI  Unnamed: 0  sell_price  delta_price_all_rel  delta_price_weekn-1  \\\n",
       "0        0         NaN         NaN                  NaN                  NaN   \n",
       "1        0         NaN         NaN                  NaN                  NaN   \n",
       "2        0         NaN         NaN                  NaN                  NaN   \n",
       "3        0         NaN         NaN                  NaN                  NaN   \n",
       "4        0         NaN         NaN                  NaN                  NaN   \n",
       "\n",
       "   delta_price_cat_rel  \n",
       "0                  NaN  \n",
       "1                  NaN  \n",
       "2                  NaN  \n",
       "3                  NaN  \n",
       "4                  NaN  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"{:,} records in df\".format(len(df)))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "del sales"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Features Engineering\n",
    "> Time Series data must be re-framed as a supervised learning dataset before we can start using machine learning algorithms:\n",
    "- There is no concept of input and output features in time series. \n",
    "- Instead, we must choose the variable to be predicted and use feature engineering to construct all of the inputs that will be used to make predictions for future time steps.\n",
    "\n",
    "### 1. Label Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Change Store id to category type\n",
    "df.store_id = df.store_id.astype('category')\n",
    "df.item_id = df.item_id.astype('category')\n",
    "df.cat_id = df.cat_id.astype('category')\n",
    "df.state_id = df.state_id.astype('category')\n",
    "df.id = df.id.astype('category')\n",
    "df.dept_id = df.dept_id.astype('category')\n",
    "\n",
    "# Store the categories along with their codes\n",
    "d_id = dict(zip(df.id.cat.codes, df.id))\n",
    "\n",
    "# Item, Department and Categories\n",
    "d_item_id = dict(zip(df.item_id.cat.codes, df.item_id))\n",
    "d_dept_id = dict(zip(df.dept_id.cat.codes, df.dept_id))\n",
    "d_cat_id = dict(zip(df.cat_id.cat.codes, df.cat_id))\n",
    "\n",
    "# Stores and States\n",
    "d_store_id = dict(zip(df.store_id.cat.codes, df.store_id))\n",
    "d_state_id = dict(zip(df.state_id.cat.codes, df.state_id))\n",
    "\n",
    "# Save to DataFrame\n",
    "LIST_SAVE = [d_id, d_item_id, d_dept_id, d_cat_id, d_store_id, d_state_id]\n",
    "LIST_NAME = ['d_id', 'd_item_id', 'd_dept_id', 'd_cat_id', 'd_store_id', 'd_state_id']\n",
    "for list_save, list_name in zip(LIST_SAVE, LIST_NAME):\n",
    "    pickle.dump(list_save, open('Data/General/{}.p'.format(list_name), \"wb\"))  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataframe shape: (60034810, 25)\n"
     ]
    },
    {
     "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>item_id</th>\n",
       "      <th>dept_id</th>\n",
       "      <th>cat_id</th>\n",
       "      <th>store_id</th>\n",
       "      <th>state_id</th>\n",
       "      <th>d</th>\n",
       "      <th>sold</th>\n",
       "      <th>wm_yr_wk</th>\n",
       "      <th>weekday</th>\n",
       "      <th>wday</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>event_name_1</th>\n",
       "      <th>event_type_1</th>\n",
       "      <th>event_name_2</th>\n",
       "      <th>event_type_2</th>\n",
       "      <th>snap_CA</th>\n",
       "      <th>snap_TX</th>\n",
       "      <th>snap_WI</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>sell_price</th>\n",
       "      <th>delta_price_all_rel</th>\n",
       "      <th>delta_price_weekn-1</th>\n",
       "      <th>delta_price_cat_rel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14370</td>\n",
       "      <td>1437</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14380</td>\n",
       "      <td>1438</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14390</td>\n",
       "      <td>1439</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14400</td>\n",
       "      <td>1440</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>14410</td>\n",
       "      <td>1441</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      id  item_id  dept_id  cat_id  store_id  state_id  d  sold  wm_yr_wk  \\\n",
       "0  14370     1437        3       1         0         0  1     0     11101   \n",
       "1  14380     1438        3       1         0         0  1     0     11101   \n",
       "2  14390     1439        3       1         0         0  1     0     11101   \n",
       "3  14400     1440        3       1         0         0  1     0     11101   \n",
       "4  14410     1441        3       1         0         0  1     0     11101   \n",
       "\n",
       "   weekday  wday  month  year  event_name_1  event_type_1  event_name_2  \\\n",
       "0        2     1      1  2011            -1            -1            -1   \n",
       "1        2     1      1  2011            -1            -1            -1   \n",
       "2        2     1      1  2011            -1            -1            -1   \n",
       "3        2     1      1  2011            -1            -1            -1   \n",
       "4        2     1      1  2011            -1            -1            -1   \n",
       "\n",
       "   event_type_2  snap_CA  snap_TX  snap_WI  Unnamed: 0  sell_price  \\\n",
       "0            -1        0        0        0         NaN         NaN   \n",
       "1            -1        0        0        0         NaN         NaN   \n",
       "2            -1        0        0        0         NaN         NaN   \n",
       "3            -1        0        0        0         NaN         NaN   \n",
       "4            -1        0        0        0         NaN         NaN   \n",
       "\n",
       "   delta_price_all_rel  delta_price_weekn-1  delta_price_cat_rel  \n",
       "0                  NaN                  NaN                  NaN  \n",
       "1                  NaN                  NaN                  NaN  \n",
       "2                  NaN                  NaN                  NaN  \n",
       "3                  NaN                  NaN                  NaN  \n",
       "4                  NaN                  NaN                  NaN  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Mapping with Category Codes\n",
    "\n",
    "# Remove d_ and transform to int (dates)\n",
    "df.d = df['d'].apply(lambda x: x.split('_')[1]).astype(np.int16)\n",
    "cols = df.dtypes.index.tolist()\n",
    "types = df.dtypes.values.tolist()\n",
    "\n",
    "# Transform categorical data to codes\n",
    "for i,type in enumerate(types):\n",
    "    if type.name == 'category':\n",
    "        df[cols[i]] = df[cols[i]].cat.codes\n",
    "        \n",
    "# Drop Dates\n",
    "df.drop('date',axis=1,inplace=True)\n",
    "print(\"Dataframe shape: {}\".format(df.shape))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Introduce Lag\n",
    "> Lag features are the classical way that time series forecasting problems are transformed into supervised learning problems.\n",
    "- Introduce lags to the the target variable sold. \n",
    "- The maximum lag I have introduced is 36 days. \n",
    "- It's purely upto you how many lags you want to introduce."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> TODO LIST: Test performances and features importance with other values of lags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>item_id</th>\n",
       "      <th>dept_id</th>\n",
       "      <th>cat_id</th>\n",
       "      <th>store_id</th>\n",
       "      <th>state_id</th>\n",
       "      <th>d</th>\n",
       "      <th>sold</th>\n",
       "      <th>wm_yr_wk</th>\n",
       "      <th>weekday</th>\n",
       "      <th>wday</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>event_name_1</th>\n",
       "      <th>event_type_1</th>\n",
       "      <th>event_name_2</th>\n",
       "      <th>event_type_2</th>\n",
       "      <th>snap_CA</th>\n",
       "      <th>snap_TX</th>\n",
       "      <th>snap_WI</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>sell_price</th>\n",
       "      <th>delta_price_all_rel</th>\n",
       "      <th>delta_price_weekn-1</th>\n",
       "      <th>delta_price_cat_rel</th>\n",
       "      <th>sold_lag_1</th>\n",
       "      <th>sold_lag_2</th>\n",
       "      <th>sold_lag_3</th>\n",
       "      <th>sold_lag_7</th>\n",
       "      <th>sold_lag_14</th>\n",
       "      <th>sold_lag_28</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14370</td>\n",
       "      <td>1437</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14380</td>\n",
       "      <td>1438</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14390</td>\n",
       "      <td>1439</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14400</td>\n",
       "      <td>1440</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>14410</td>\n",
       "      <td>1441</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11101</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      id  item_id  dept_id  cat_id  store_id  state_id  d  sold  wm_yr_wk  \\\n",
       "0  14370     1437        3       1         0         0  1     0     11101   \n",
       "1  14380     1438        3       1         0         0  1     0     11101   \n",
       "2  14390     1439        3       1         0         0  1     0     11101   \n",
       "3  14400     1440        3       1         0         0  1     0     11101   \n",
       "4  14410     1441        3       1         0         0  1     0     11101   \n",
       "\n",
       "   weekday  wday  month  year  event_name_1  event_type_1  event_name_2  \\\n",
       "0        2     1      1  2011            -1            -1            -1   \n",
       "1        2     1      1  2011            -1            -1            -1   \n",
       "2        2     1      1  2011            -1            -1            -1   \n",
       "3        2     1      1  2011            -1            -1            -1   \n",
       "4        2     1      1  2011            -1            -1            -1   \n",
       "\n",
       "   event_type_2  snap_CA  snap_TX  snap_WI  Unnamed: 0  sell_price  \\\n",
       "0            -1        0        0        0         NaN         NaN   \n",
       "1            -1        0        0        0         NaN         NaN   \n",
       "2            -1        0        0        0         NaN         NaN   \n",
       "3            -1        0        0        0         NaN         NaN   \n",
       "4            -1        0        0        0         NaN         NaN   \n",
       "\n",
       "   delta_price_all_rel  delta_price_weekn-1  delta_price_cat_rel  sold_lag_1  \\\n",
       "0                  NaN                  NaN                  NaN         NaN   \n",
       "1                  NaN                  NaN                  NaN         NaN   \n",
       "2                  NaN                  NaN                  NaN         NaN   \n",
       "3                  NaN                  NaN                  NaN         NaN   \n",
       "4                  NaN                  NaN                  NaN         NaN   \n",
       "\n",
       "   sold_lag_2  sold_lag_3  sold_lag_7  sold_lag_14  sold_lag_28  \n",
       "0         NaN         NaN         NaN          NaN          NaN  \n",
       "1         NaN         NaN         NaN          NaN          NaN  \n",
       "2         NaN         NaN         NaN          NaN          NaN  \n",
       "3         NaN         NaN         NaN          NaN          NaN  \n",
       "4         NaN         NaN         NaN          NaN          NaN  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Introduce lags (days)\n",
    "lags = [1, 2, 3, 7, 14, 28]\n",
    "for lag in lags:\n",
    "    df['sold_lag_'+str(lag)] = df.groupby(\n",
    "        ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'], \n",
    "        as_index = False)['sold'].shift(lag).astype(np.float16)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Mean Encoding\n",
    "> From a mathematical point of view, mean encoding represents a probability of your target variable, conditional on each value of the feature. In a way, it embodies the target variable in its encoded value. I have calculated mean encodings on the basis of following logical features\n",
    "\n",
    "- item\n",
    "- state\n",
    "- store\n",
    "- category\n",
    "- department\n",
    "- category & department\n",
    "- store & item\n",
    "- category & item\n",
    "- department & item\n",
    "- state & store\n",
    "- state, store and category\n",
    "- store, category and department"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Total Average Sales by: item, state, store, cat and dept\n",
    "df['iteam_sold_avg'] = df.groupby('item_id')['sold'].transform('mean').astype(np.float16)\n",
    "df['state_sold_avg'] = df.groupby('state_id')['sold'].transform('mean').astype(np.float16)\n",
    "df['store_sold_avg'] = df.groupby('store_id')['sold'].transform('mean').astype(np.float16)\n",
    "df['cat_sold_avg'] = df.groupby('cat_id')['sold'].transform('mean').astype(np.float16)\n",
    "df['dept_sold_avg'] = df.groupby('dept_id')['sold'].transform('mean').astype(np.float16)\n",
    "\n",
    "# Sales average by \n",
    "df['cat_dept_sold_avg'] = df.groupby(['cat_id','dept_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['store_item_sold_avg'] = df.groupby(['store_id','item_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['cat_item_sold_avg'] = df.groupby(['cat_id','item_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['dept_item_sold_avg'] = df.groupby(['dept_id','item_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['state_store_sold_avg'] = df.groupby(['state_id','store_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['state_store_cat_sold_avg'] = df.groupby(['state_id','store_id','cat_id'])['sold'].transform('mean').astype(np.float16)\n",
    "df['store_cat_dept_sold_avg'] = df.groupby(['store_id','cat_id','dept_id'])['sold'].transform('mean').astype(np.float16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Rolling Average (Window) on actual sales\n",
    "| Days = 7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df['rolling_sold_mean'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold'].transform(\n",
    "    lambda x: x.rolling(window=7).mean()).astype(np.float16)\n",
    "\n",
    "# Average for the last n days\n",
    "for days in [3, 7, 14, 21, 28]:\n",
    "    df['rolling_sold_mean_{}'.format(days)] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold'].transform(\n",
    "    lambda x: x.rolling(window=days).mean()).astype(np.float16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Rolling Average (Window) on actual lag\n",
    "- rmean_7_7: rolling mean sales of a window size of 7 over column lag_7\n",
    "- rmean_7_28: rolling mean sales of a window size of 7 over column lag_28\n",
    "- rmean_28_7: rolling mean sales of a window size of 28 over column lag_7\n",
    "- rmean_28_28: rolling mean sales of a window size of 28 over column lag_28"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Rolling Average on actual lag\n",
    "for window, lag in zip([7, 7, 28, 28], [7, 28, 7, 28]):\n",
    "    df['rolling_lag_{}_win_{}'.format(window, lag)] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold_lag_{}'.format(lag)].transform(\n",
    "    lambda x: x.rolling(window=window).mean()).astype(np.float16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Trends \n",
    "By SKU (Store + Item Id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Daily Average by SKU (Item Id + Store)\n",
    "df['daily_avg_sold'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id','d'])['sold'].transform('mean').astype(np.float16)\n",
    "# Total Average by SKU (Item Id + Store)\n",
    "df['avg_sold'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold'].transform('mean').astype(np.float16)\n",
    "# Selling Trend\n",
    "df['selling_trend'] = (df['daily_avg_sold'] - df['avg_sold']).astype(np.float16)\n",
    "\n",
    "# Drop Columns \n",
    "df.drop(['daily_avg_sold','avg_sold'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## /!\\ Additional Ideas for Features Engineering /!\\ \n",
    "### Trends by SKU for all stores\n",
    "By Item (All Stores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Daily Average by SKU (Item Id + Store)\n",
    "df['item_daily_avg_sold'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'd'])['sold'].transform('mean').astype(np.float16)\n",
    "# Total Average by SKU (Item Id + Store)\n",
    "df['item_avg_sold'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id'])['sold'].transform('mean').astype(np.float16)\n",
    "# Selling Trend\n",
    "df['item_selling_trend'] = (df['item_daily_avg_sold'] - df['item_avg_sold']).astype(np.float16)\n",
    "\n",
    "# Drop Columns \n",
    "df.drop(['item_daily_avg_sold','item_avg_sold'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rolling Average (Window) on actual lag\n",
    "- rmean_7_7: rolling mean sales of a window size of 7 over column lag_7\n",
    "- rmean_7_28: rolling mean sales of a window size of 7 over column lag_28\n",
    "- rmean_28_7: rolling mean sales of a window size of 28 over column lag_7\n",
    "- rmean_28_28: rolling mean sales of a window size of 28 over column lag_28\n",
    "\n",
    "Insipiration: https://www.kaggle.com/kneroma/m5-first-public-notebook-under-0-50/notebook\n",
    "> Great explanation in comments based on business insights (why 7, 28 days?)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Rolling Average on actual lag\n",
    "for window, lag in zip([7, 7, 28, 28], [7, 28, 7, 28]):\n",
    "    df['rolling_lag_{}_win_{}'.format(window, lag)] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold_lag_{}'.format(lag)].transform(\n",
    "    lambda x: x.rolling(window=window).mean()).astype(np.float16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rolling Max (Window): Last n days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df['rolling_sold_max'] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold'].transform(\n",
    "    lambda x: x.rolling(window=7).max()).astype(np.float16)\n",
    "\n",
    "# Average for the last n days\n",
    "for days in [1, 2, 7, 14, 21, 28]:\n",
    "    df['rolling_sold_max_{}'.format(days)] = df.groupby(\n",
    "    ['id', 'item_id', 'dept_id', 'cat_id', 'store_id', 'state_id'])['sold'].transform(\n",
    "    lambda x: x.rolling(window=days).max()).astype(np.float16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stock-Outs: Do you have stock availability issue that cause zero sales in the last n days ?\n",
    "\n",
    "Definition: \n",
    "- When the inventory level is below the customer demand we have a stock-out (i.e if you have 0 units in your shelves and customer are looking for your products => stock-out)\n",
    "\n",
    "> Assumption: we will suppose that if you do not have sales for 3 days in a row you have a stock-out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30,490 unique id with 30,489.0 id changes\n"
     ]
    }
   ],
   "source": [
    "# Sort the dataset by item_id and day\n",
    "df.sort_values(['id', 'd'], ascending =[True, True], inplace = True)\n",
    "\n",
    "# Mapping id change\n",
    "df['id_change'] = df['id'].diff().fillna(0)\n",
    "print(\"{:,} unique id with {:,} id changes\".format(df.id.nunique(), df['id_change'].sum()))\n",
    "# Zero Sale \n",
    "df['zero_sale'] = (df['sold'] == 0) * (df['id_change']==False) * 1\n",
    "df['stock_out_id'] = (df['zero_sale'].transform(\n",
    "        lambda x: x.rolling(window=3).sum()).astype(np.float16)==3) * 1\n",
    "\n",
    "# Drop useless columns\n",
    "df.drop(['id_change', 'zero_sale'], axis = 1, inplace = True)\n",
    "\n",
    "# Stock-Outs in the last n days ?\n",
    "for n_days in [1, 2, 7]: \n",
    "    df['stock_out_id_last_{}_days'.format(n_days)] = (df['stock_out_id'].transform(\n",
    "        lambda x: x.rolling(window=n_days).sum()).astype(np.float16)>0) * 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Store Opening: Was the store closed in the last n days ?\n",
    "\n",
    "> Assumption: if the store total sales is zero => closed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Store Closed = Sales zero\n",
    "df['store_closed'] = (df.groupby(\n",
    "    ['store_id', 'd'])['sold'].transform('sum').astype(np.float16) == 0) * 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Store Closed = Sales zero\n",
    "for n_days in [1, 2, 7]: # closed the last week \n",
    "    df['store_closed_last_{}_days'.format(n_days)] = (df['store_closed'].transform(\n",
    "        lambda x: x.rolling(window=n_days).sum()).astype(np.float16)>0) * 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df.to_pickle('Data/data_features_improved.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "del df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modelling and Prediction\n",
    "> Model = LGBM\n",
    "- Choose features (original + additional)\n",
    "- TO DO Search Grid (I can't because of my limited computing ressources :))\n",
    "- Measure Performance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import Procesed DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "58,388,350 records for the prepared data\n"
     ]
    },
    {
     "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>item_id</th>\n",
       "      <th>dept_id</th>\n",
       "      <th>cat_id</th>\n",
       "      <th>store_id</th>\n",
       "      <th>state_id</th>\n",
       "      <th>d</th>\n",
       "      <th>sold</th>\n",
       "      <th>wm_yr_wk</th>\n",
       "      <th>weekday</th>\n",
       "      <th>wday</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>event_name_1</th>\n",
       "      <th>event_type_1</th>\n",
       "      <th>event_name_2</th>\n",
       "      <th>event_type_2</th>\n",
       "      <th>snap_CA</th>\n",
       "      <th>snap_TX</th>\n",
       "      <th>snap_WI</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>sell_price</th>\n",
       "      <th>delta_price_all_rel</th>\n",
       "      <th>delta_price_weekn-1</th>\n",
       "      <th>delta_price_cat_rel</th>\n",
       "      <th>sold_lag_1</th>\n",
       "      <th>sold_lag_2</th>\n",
       "      <th>sold_lag_3</th>\n",
       "      <th>sold_lag_7</th>\n",
       "      <th>sold_lag_14</th>\n",
       "      <th>sold_lag_28</th>\n",
       "      <th>iteam_sold_avg</th>\n",
       "      <th>state_sold_avg</th>\n",
       "      <th>store_sold_avg</th>\n",
       "      <th>cat_sold_avg</th>\n",
       "      <th>dept_sold_avg</th>\n",
       "      <th>cat_dept_sold_avg</th>\n",
       "      <th>store_item_sold_avg</th>\n",
       "      <th>cat_item_sold_avg</th>\n",
       "      <th>dept_item_sold_avg</th>\n",
       "      <th>state_store_sold_avg</th>\n",
       "      <th>state_store_cat_sold_avg</th>\n",
       "      <th>store_cat_dept_sold_avg</th>\n",
       "      <th>rolling_sold_mean</th>\n",
       "      <th>rolling_sold_mean_3</th>\n",
       "      <th>rolling_sold_mean_7</th>\n",
       "      <th>rolling_sold_mean_14</th>\n",
       "      <th>rolling_sold_mean_21</th>\n",
       "      <th>rolling_sold_mean_28</th>\n",
       "      <th>rolling_lag_7_win_7</th>\n",
       "      <th>rolling_lag_7_win_28</th>\n",
       "      <th>rolling_lag_28_win_7</th>\n",
       "      <th>rolling_lag_28_win_28</th>\n",
       "      <th>selling_trend</th>\n",
       "      <th>item_selling_trend</th>\n",
       "      <th>rolling_sold_max</th>\n",
       "      <th>rolling_sold_max_1</th>\n",
       "      <th>rolling_sold_max_2</th>\n",
       "      <th>rolling_sold_max_7</th>\n",
       "      <th>rolling_sold_max_14</th>\n",
       "      <th>rolling_sold_max_21</th>\n",
       "      <th>rolling_sold_max_28</th>\n",
       "      <th>stock_out_id</th>\n",
       "      <th>stock_out_id_last_1_days</th>\n",
       "      <th>stock_out_id_last_2_days</th>\n",
       "      <th>stock_out_id_last_7_days</th>\n",
       "      <th>store_closed</th>\n",
       "      <th>store_closed_last_1_days</th>\n",
       "      <th>store_closed_last_2_days</th>\n",
       "      <th>store_closed_last_7_days</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1648072</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>11108</td>\n",
       "      <td>4</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>368753.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.322754</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.21582</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.623047</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>0.774902</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.933594</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>0.714355</td>\n",
       "      <td>0.666504</td>\n",
       "      <td>0.714355</td>\n",
       "      <td>0.856934</td>\n",
       "      <td>1.142578</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.571289</td>\n",
       "      <td>1.713867</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.774902</td>\n",
       "      <td>-0.774902</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1678562</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>1</td>\n",
       "      <td>11108</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>368753.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.322754</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.21582</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.623047</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>0.774902</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.933594</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>0.856934</td>\n",
       "      <td>0.666504</td>\n",
       "      <td>0.856934</td>\n",
       "      <td>0.928711</td>\n",
       "      <td>1.142578</td>\n",
       "      <td>1.392578</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.857422</td>\n",
       "      <td>1.642578</td>\n",
       "      <td>1.392578</td>\n",
       "      <td>0.225098</td>\n",
       "      <td>0.225098</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1709052</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>57</td>\n",
       "      <td>2</td>\n",
       "      <td>11109</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>368754.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.322754</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.21582</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.623047</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>0.774902</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.933594</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.142578</td>\n",
       "      <td>1.392578</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.642578</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>1.224609</td>\n",
       "      <td>1.224609</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1739542</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>58</td>\n",
       "      <td>2</td>\n",
       "      <td>11109</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>368754.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.322754</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.21582</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.623047</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>0.774902</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.933594</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>1.142578</td>\n",
       "      <td>1.666992</td>\n",
       "      <td>1.142578</td>\n",
       "      <td>0.785645</td>\n",
       "      <td>1.095703</td>\n",
       "      <td>1.392578</td>\n",
       "      <td>0.428467</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.607422</td>\n",
       "      <td>1.428711</td>\n",
       "      <td>1.224609</td>\n",
       "      <td>1.224609</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1770032</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>11109</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>368754.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.322754</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.21582</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.623047</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>1.220703</td>\n",
       "      <td>0.774902</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>0.629883</td>\n",
       "      <td>1.304688</td>\n",
       "      <td>1.933594</td>\n",
       "      <td>1.357422</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.333008</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.785645</td>\n",
       "      <td>1.095703</td>\n",
       "      <td>1.392578</td>\n",
       "      <td>0.571289</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.642578</td>\n",
       "      <td>1.428711</td>\n",
       "      <td>-0.774902</td>\n",
       "      <td>-0.774902</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         id  item_id  dept_id  cat_id  store_id  state_id   d  sold  wm_yr_wk  \\\n",
       "1648072   0        0        0       0         0         0  55     0     11108   \n",
       "1678562   0        0        0       0         0         0  56     1     11108   \n",
       "1709052   0        0        0       0         0         0  57     2     11109   \n",
       "1739542   0        0        0       0         0         0  58     2     11109   \n",
       "1770032   0        0        0       0         0         0  59     0     11109   \n",
       "\n",
       "         weekday  wday  month  year  event_name_1  event_type_1  event_name_2  \\\n",
       "1648072        4     6      3  2011            -1            -1            -1   \n",
       "1678562        0     7      3  2011            -1            -1            -1   \n",
       "1709052        2     1      3  2011            -1            -1            -1   \n",
       "1739542        3     2      3  2011            -1            -1            -1   \n",
       "1770032        1     3      3  2011            -1            -1            -1   \n",
       "\n",
       "         event_type_2  snap_CA  snap_TX  snap_WI  Unnamed: 0  sell_price  \\\n",
       "1648072            -1        0        0        0    368753.0         2.0   \n",
       "1678562            -1        0        0        0    368753.0         2.0   \n",
       "1709052            -1        0        0        0    368754.0         2.0   \n",
       "1739542            -1        0        0        0    368754.0         2.0   \n",
       "1770032            -1        0        0        0    368754.0         2.0   \n",
       "\n",
       "         delta_price_all_rel  delta_price_weekn-1  delta_price_cat_rel  \\\n",
       "1648072                  0.0                  0.0            -0.322754   \n",
       "1678562                  0.0                  0.0            -0.322754   \n",
       "1709052                  0.0                  0.0            -0.322754   \n",
       "1739542                  0.0                  0.0            -0.322754   \n",
       "1770032                  0.0                  0.0            -0.322754   \n",
       "\n",
       "         sold_lag_1  sold_lag_2  sold_lag_3  sold_lag_7  sold_lag_14  \\\n",
       "1648072         1.0         1.0         1.0         0.0          3.0   \n",
       "1678562         0.0         1.0         1.0         0.0          0.0   \n",
       "1709052         1.0         0.0         1.0         1.0          1.0   \n",
       "1739542         2.0         1.0         0.0         1.0          5.0   \n",
       "1770032         2.0         2.0         1.0         1.0          0.0   \n",
       "\n",
       "         sold_lag_28  iteam_sold_avg  state_sold_avg  store_sold_avg  \\\n",
       "1648072          2.0        0.629883         1.21582        1.304688   \n",
       "1678562          4.0        0.629883         1.21582        1.304688   \n",
       "1709052          2.0        0.629883         1.21582        1.304688   \n",
       "1739542          2.0        0.629883         1.21582        1.304688   \n",
       "1770032          0.0        0.629883         1.21582        1.304688   \n",
       "\n",
       "         cat_sold_avg  dept_sold_avg  cat_dept_sold_avg  store_item_sold_avg  \\\n",
       "1648072      1.623047       1.220703           1.220703             0.774902   \n",
       "1678562      1.623047       1.220703           1.220703             0.774902   \n",
       "1709052      1.623047       1.220703           1.220703             0.774902   \n",
       "1739542      1.623047       1.220703           1.220703             0.774902   \n",
       "1770032      1.623047       1.220703           1.220703             0.774902   \n",
       "\n",
       "         cat_item_sold_avg  dept_item_sold_avg  state_store_sold_avg  \\\n",
       "1648072           0.629883            0.629883              1.304688   \n",
       "1678562           0.629883            0.629883              1.304688   \n",
       "1709052           0.629883            0.629883              1.304688   \n",
       "1739542           0.629883            0.629883              1.304688   \n",
       "1770032           0.629883            0.629883              1.304688   \n",
       "\n",
       "         state_store_cat_sold_avg  store_cat_dept_sold_avg  rolling_sold_mean  \\\n",
       "1648072                  1.933594                 1.357422           0.714355   \n",
       "1678562                  1.933594                 1.357422           0.856934   \n",
       "1709052                  1.933594                 1.357422           1.000000   \n",
       "1739542                  1.933594                 1.357422           1.142578   \n",
       "1770032                  1.933594                 1.357422           1.000000   \n",
       "\n",
       "         rolling_sold_mean_3  rolling_sold_mean_7  rolling_sold_mean_14  \\\n",
       "1648072             0.666504             0.714355              0.856934   \n",
       "1678562             0.666504             0.856934              0.928711   \n",
       "1709052             1.000000             1.000000              1.000000   \n",
       "1739542             1.666992             1.142578              0.785645   \n",
       "1770032             1.333008             1.000000              0.785645   \n",
       "\n",
       "         rolling_sold_mean_21  rolling_sold_mean_28  rolling_lag_7_win_7  \\\n",
       "1648072              1.142578              1.500000             1.000000   \n",
       "1678562              1.142578              1.392578             1.000000   \n",
       "1709052              1.142578              1.392578             1.000000   \n",
       "1739542              1.095703              1.392578             0.428467   \n",
       "1770032              1.095703              1.392578             0.571289   \n",
       "\n",
       "         rolling_lag_7_win_28  rolling_lag_28_win_7  rolling_lag_28_win_28  \\\n",
       "1648072              1.571289              1.713867                    NaN   \n",
       "1678562              1.857422              1.642578               1.392578   \n",
       "1709052              2.000000              1.642578               1.357422   \n",
       "1739542              2.000000              1.607422               1.428711   \n",
       "1770032              2.000000              1.642578               1.428711   \n",
       "\n",
       "         selling_trend  item_selling_trend  rolling_sold_max  \\\n",
       "1648072      -0.774902           -0.774902               1.0   \n",
       "1678562       0.225098            0.225098               1.0   \n",
       "1709052       1.224609            1.224609               2.0   \n",
       "1739542       1.224609            1.224609               2.0   \n",
       "1770032      -0.774902           -0.774902               2.0   \n",
       "\n",
       "         rolling_sold_max_1  rolling_sold_max_2  rolling_sold_max_7  \\\n",
       "1648072                 0.0                 1.0                 1.0   \n",
       "1678562                 1.0                 1.0                 1.0   \n",
       "1709052                 2.0                 2.0                 2.0   \n",
       "1739542                 2.0                 2.0                 2.0   \n",
       "1770032                 0.0                 2.0                 2.0   \n",
       "\n",
       "         rolling_sold_max_14  rolling_sold_max_21  rolling_sold_max_28  \\\n",
       "1648072                  5.0                  5.0                  7.0   \n",
       "1678562                  5.0                  5.0                  7.0   \n",
       "1709052                  5.0                  5.0                  7.0   \n",
       "1739542                  2.0                  5.0                  7.0   \n",
       "1770032                  2.0                  5.0                  7.0   \n",
       "\n",
       "         stock_out_id  stock_out_id_last_1_days  stock_out_id_last_2_days  \\\n",
       "1648072             0                         0                         0   \n",
       "1678562             0                         0                         0   \n",
       "1709052             0                         0                         0   \n",
       "1739542             0                         0                         0   \n",
       "1770032             0                         0                         0   \n",
       "\n",
       "         stock_out_id_last_7_days  store_closed  store_closed_last_1_days  \\\n",
       "1648072                         0             0                         0   \n",
       "1678562                         0             0                         0   \n",
       "1709052                         0             0                         0   \n",
       "1739542                         0             0                         0   \n",
       "1770032                         0             0                         0   \n",
       "\n",
       "         store_closed_last_2_days  store_closed_last_7_days  \n",
       "1648072                         0                         0  \n",
       "1678562                         0                         0  \n",
       "1709052                         0                         0  \n",
       "1739542                         0                         0  \n",
       "1770032                         0                         0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Import Data\n",
    "data_size = 'improved'\n",
    "data = pd.read_pickle('Data/data_features_{}.pkl'.format(data_size))\n",
    "\n",
    "# Start records from 55th\n",
    "start_row = 55\n",
    "data = data[data['d'] >= start_row].copy()\n",
    "print(\"{:,} records for the prepared data\".format(len(data)))\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import key data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "LIST_NAME = ['d_id', 'd_item_id', 'd_dept_id', 'd_cat_id', 'd_store_id', 'd_state_id']\n",
    "dict_data = {}\n",
    "for list_name in LIST_NAME:\n",
    "    dict_temp = pickle.load(open('Data/General/{}.p'.format(list_name), \"rb\"))\n",
    "    dict_data[list_name] = dict_temp\n",
    "    del dict_temp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Show Features\n",
    "> We'll split the features we created in groups so we can add them progressively and measure the impact on the accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Initial Features\n",
    "INIT_FEAT = list(data.columns[0:21])\n",
    "\n",
    "# Lags and averages\n",
    "LAGAV_FEAT = list(data.columns[24:42])\n",
    "\n",
    "# Rolling Means and Rolling Means on lag\n",
    "ROLLMEAN_FEAT = list(data.columns[42:52])\n",
    "\n",
    "# Trends and Rolling MAX\n",
    "TREND_MAX_FEAT = list(data.columns[52:58])\n",
    "\n",
    "# Stock-Out and Store Closed\n",
    "# SO_CLOSED_FEAT = list(data.columns[58:67])\n",
    "SO_CLOSED_FEAT = list(['stock_out_id', 'store_closed'])\n",
    "\n",
    "# PRICE COMPARISON\n",
    "PRICE_COMPARE = list(data.columns[21:24])\n",
    "\n",
    "# Dictionnary with different steps\n",
    "dict_features = {\n",
    "    'STEP_1': INIT_FEAT,\n",
    "    'STEP_2': INIT_FEAT+LAGAV_FEAT,\n",
    "    'STEP_3': INIT_FEAT+LAGAV_FEAT+ROLLMEAN_FEAT,\n",
    "    'STEP_4': INIT_FEAT+LAGAV_FEAT+ROLLMEAN_FEAT+TREND_MAX_FEAT,\n",
    "    'STEP_5': INIT_FEAT+LAGAV_FEAT+ROLLMEAN_FEAT+TREND_MAX_FEAT+SO_CLOSED_FEAT,\n",
    "    'STEP_6': INIT_FEAT+LAGAV_FEAT+ROLLMEAN_FEAT+TREND_MAX_FEAT+SO_CLOSED_FEAT+PRICE_COMPARE,\n",
    "}\n",
    "LIST_STEPS = ['STEP_1', 'STEP_2', 'STEP_3', 'STEP_4', 'STEP_5', 'STEP_6']\n",
    "LIST_STEPS_NAME = ['INITIAL_DATA', 'INITIAL + LAG + AVERAGES', \n",
    "                  'INITIAL + LAG + AVERAGES + ROLLING MEAN',\n",
    "                  'INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX',\n",
    "                  'INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED',\n",
    "                  'INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED + PRICE COMPARISON']\n",
    "dict_stepname = dict(zip(LIST_STEPS, LIST_STEPS_NAME))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Train and Test Models\n",
    "#### Create Validation and Test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Validation Set\n",
    "valid = data[(data['d']>=1914) & (data['d']<1942)][['id','d','sold']]\n",
    "# Validation Prediction\n",
    "valid_preds = valid['sold']\n",
    "\n",
    "# Test Set\n",
    "test = data[data['d'] >= 1942][['id', 'd', 'sold']]\n",
    "# Evaluation Prediction\n",
    "eval_preds = test['sold']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train and Test Models with different features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prediction for STEP: INITIAL_DATA*****\n",
      "For INITIAL_DATA: RMSE = 166.5922878247231\n",
      "*****Prediction for STEP: INITIAL + LAG + AVERAGES*****\n",
      "For INITIAL + LAG + AVERAGES: RMSE = 136.86106334740734\n",
      "*****Prediction for STEP: INITIAL + LAG + AVERAGES + ROLLING MEAN*****\n",
      "For INITIAL + LAG + AVERAGES + ROLLING MEAN: RMSE = 18.613120309842593\n",
      "*****Prediction for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX*****\n",
      "For INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX: RMSE = 6.47061859337155\n",
      "*****Prediction for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED*****\n",
      "For INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED: RMSE = 6.355151609638661\n",
      "*****Prediction for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED + PRICE COMPARISON*****\n",
      "For INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED + PRICE COMPARISON: RMSE = 4.598720935658507\n"
     ]
    },
    {
     "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>STEP</th>\n",
       "      <th>STEP_NAME</th>\n",
       "      <th>rmse</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>STEP_1</td>\n",
       "      <td>INITIAL_DATA</td>\n",
       "      <td>166.592288</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>STEP_2</td>\n",
       "      <td>INITIAL + LAG + AVERAGES</td>\n",
       "      <td>136.861063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>STEP_3</td>\n",
       "      <td>INITIAL + LAG + AVERAGES + ROLLING MEAN</td>\n",
       "      <td>18.613120</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>STEP_4</td>\n",
       "      <td>INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...</td>\n",
       "      <td>6.470619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>STEP_5</td>\n",
       "      <td>INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...</td>\n",
       "      <td>6.355152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>STEP_6</td>\n",
       "      <td>INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...</td>\n",
       "      <td>4.598721</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     STEP                                          STEP_NAME        rmse\n",
       "0  STEP_1                                       INITIAL_DATA  166.592288\n",
       "1  STEP_2                           INITIAL + LAG + AVERAGES  136.861063\n",
       "2  STEP_3            INITIAL + LAG + AVERAGES + ROLLING MEAN   18.613120\n",
       "3  STEP_4  INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...    6.470619\n",
       "4  STEP_5  INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...    6.355152\n",
       "5  STEP_6  INITIAL + LAG AND AVERAGES + ROLLING MEAN + TR...    4.598721"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get the store ids\n",
    "stores = data.store_id.unique()\n",
    "d_store_id = dict_data['d_store_id']\n",
    "\n",
    "# Dictionnary with errors each step\n",
    "dict_error = {}\n",
    "\n",
    "# Loop with the steps\n",
    "for step in LIST_STEPS:\n",
    "\n",
    "    # Folder for Models \n",
    "    print('*****Prediction for STEP: {}*****'.format(dict_stepname[step]))\n",
    "    FOLDER_MODEL = 'Model/{}/{}_Features_Improved/'.format(data_size, step)\n",
    "    Path(FOLDER_MODEL).mkdir(parents=True, exist_ok=True)\n",
    "    COLS_SCOPE = dict_features[step]\n",
    "\n",
    "    # DataFrame with filter scope\n",
    "    data_scope = data[COLS_SCOPE].copy()\n",
    "\n",
    "    # Validation Set\n",
    "    valid = data_scope[(data_scope['d']>=1914) & (data_scope['d']<1942)][['id','d','sold']]\n",
    "    # Validation Prediction\n",
    "    valid_set = valid['sold']\n",
    "\n",
    "    # Validation + Predicition for all stores by step \n",
    "    df_validpred = pd.DataFrame()\n",
    "\n",
    "    # Loop for training a model for each store\n",
    "    for store in stores:\n",
    "\n",
    "        # Dataframe for each store\n",
    "        df = data_scope[data_scope['store_id']==store]\n",
    "\n",
    "        # Train Data until day = 1914\n",
    "        X_train, y_train = df[df['d']<1914].drop('sold',axis=1), df[df['d']<1914]['sold']\n",
    "\n",
    "        # Validation Day: 1914 to 1942\n",
    "        X_valid, y_valid = df[(df['d']>=1914) & (df['d']<1942)].drop('sold',axis=1), df[(df['d']>=1914) & (df['d']<1942)]['sold']\n",
    "\n",
    "        # X_test with \n",
    "        X_test = df[df['d']>=1942].drop('sold',axis=1)\n",
    "        \n",
    "        # Train and validate\n",
    "        model = LGBMRegressor(\n",
    "            n_estimators=1000,\n",
    "            learning_rate=0.3,\n",
    "            subsample=0.8,\n",
    "            colsample_bytree=0.8,\n",
    "            max_depth=8,\n",
    "            num_leaves=50,\n",
    "            min_child_weight=300\n",
    "        )\n",
    "\n",
    "\n",
    "        # Fit model\n",
    "        model.fit(X_train, y_train, \n",
    "                  eval_set = [(X_train,y_train),(X_valid,y_valid)], \n",
    "                  eval_metric = 'rmse', \n",
    "                  verbose = 0, \n",
    "                  early_stopping_rounds = 20)\n",
    "\n",
    "        # Compute Prediction\n",
    "        valid_pred = model.predict(X_valid)\n",
    "\n",
    "        # Actual Validation vs. Prediction\n",
    "        df_valid = pd.DataFrame({\n",
    "            'validation':valid_set[X_valid.index],\n",
    "            'prediction':valid_pred,\n",
    "            'store': d_store_id[store]\n",
    "        })\n",
    "        df_valid['error'] = df_valid['validation'] - df_valid['prediction']\n",
    "        df_validpred = pd.concat([df_validpred, df_valid])\n",
    "        # Save prediction\n",
    "        df_valid.to_csv(FOLDER_MODEL + 'prediction_{}.csv'.format(step))\n",
    "\n",
    "        # Save model\n",
    "        filename = FOLDER_MODEL + 'model_features_total-' + str(d_store_id[store])+'.pkl'\n",
    "        joblib.dump(model, filename)\n",
    "\n",
    "        del model, X_train, y_train, X_valid, y_valid\n",
    "\n",
    "    # Save Prediction for all stores\n",
    "    df_validpred.to_csv(FOLDER_MODEL + 'prediction_{}.csv'.format(step))\n",
    "\n",
    "    # Compute Error\n",
    "    valid_rmse = 100 * np.sqrt(np.mean((df_validpred.validation.values - df_validpred.prediction.values) ** 2))/np.mean(df_validpred.validation.values)\n",
    "    # Add Error in a Dictionnary\n",
    "    dict_error[step] = valid_rmse\n",
    "    print(\"For {}: RMSE = {}\".format(dict_stepname[step], valid_rmse))\n",
    "\n",
    "# Final DataFrame with error for all stores by STEP\n",
    "df_error = pd.DataFrame({\n",
    "    'STEP': LIST_STEPS,\n",
    "    'STEP_NAME': [dict_stepname[step] for step in LIST_STEPS],\n",
    "    'rmse': [dict_error[step] for step in LIST_STEPS]\n",
    "    })\n",
    "df_error.to_excel(FOLDER_MODEL + 'df_error.xlsx')\n",
    "df_error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtMAAAIKCAYAAADhxcPWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nYII=\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot\n",
    "ax = plt.gca()\n",
    "df_error.plot.bar(figsize=(12, 8), x='STEP', y='rmse', ax=ax, color = 'black')\n",
    "plt.xlabel('Features')\n",
    "plt.ylabel('RMSE (%)')\n",
    "plt.title('Root Mean Square Error (%) for each step of the experiment')\n",
    "plt.xticks(rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Additional Analysis\n",
    "### Plot Features Importance for each model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL_DATA*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAALICAYAAADFbA85AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nV78wwjOpHx8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL + LAG + AVERAGES*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAALICAYAAAAqrmQHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nLEsOKDMzy9L/B6XHR77Tm44VAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL + LAG + AVERAGES + ROLLING MEAN*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAALICAYAAAAqrmQHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAALICAYAAADFbA85AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nFQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAALICAYAAAAqrmQHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nRK5CYII=\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*****Prot Features Importance for STEP: INITIAL + LAG AND AVERAGES + ROLLING MEAN + TREND AND ROLLING MAX + STOCK-OUT AND STORE CLOSED + PRICE COMPARISON*****\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAALICAYAAAAqrmQHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90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\nRK5CYII=\n",
      "text/plain": [
       "<Figure size 432x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Folder for Models \n",
    "for step in LIST_STEPS:\n",
    "    print('*****Prot Features Importance for STEP: {}*****'.format(dict_stepname[step]))\n",
    "\n",
    "    FOLDER_MODEL = 'Model/{}/{}_Features_Improved/'.format(data_size, step)\n",
    "    Path(FOLDER_MODEL).mkdir(parents=True, exist_ok=True)\n",
    "    COLS_SCOPE = dict_features[step]\n",
    "\n",
    "    feature_importance_df = pd.DataFrame()\n",
    "    features = [f for f in COLS_SCOPE if f != 'sold']\n",
    "\n",
    "    for store in stores:\n",
    "        store_name = d_store_id[store]\n",
    "        filename = FOLDER_MODEL + 'model_features_total-' + str(d_store_id[store])+'.pkl'\n",
    "        # load model\n",
    "        model = joblib.load(filename)\n",
    "\n",
    "        # Create features importance for this model\n",
    "        store_importance_df = pd.DataFrame()\n",
    "        store_importance_df[\"feature\"] = features\n",
    "        store_importance_df[\"importance\"] = model.feature_importances_\n",
    "        store_importance_df[\"store\"] = store_name\n",
    "\n",
    "        # Concat\n",
    "        feature_importance_df = pd.concat([feature_importance_df, store_importance_df], axis=0)\n",
    "\n",
    "    # Features Importances Analysis\n",
    "    df_fi_mean = pd.DataFrame(feature_importance_df[\n",
    "        [\"feature\", \"importance\"]].groupby(\"feature\").mean().sort_values(\n",
    "        by=\"importance\", ascending=False))\n",
    "    df_fi_mean.columns = ['importance']\n",
    "    df_fi_mean['%_importance'] = (100 * df_fi_mean['importance']/df_fi_mean['importance'].sum(axis =0)).round(2)\n",
    "    df_fi_mean.to_excel(FOLDER_MODEL + 'Features_Importance_{}.xlsx'.format(step))\n",
    "\n",
    "\n",
    "    # Plot\n",
    "    cols = df_fi_mean[:20].index\n",
    "    best_features = feature_importance_df.loc[feature_importance_df.feature.isin(cols)]\n",
    "\n",
    "    fig = plt.figure(figsize=(6,10))\n",
    "    ax = fig.gca()\n",
    "    sns.barplot(x=\"importance\", y=\"feature\", data = best_features.sort_values(by=\"importance\", ascending=False))\n",
    "    plt.title('LightGBM Top 20 Features: {}'.format(step))\n",
    "    plt.tight_layout()\n",
    "    fig.savefig(FOLDER_MODEL + 'Features_Importance_{}.png'.format(step))\n",
    "    plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}