final_notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Final Assignment MLiM - WS 20/21\n",
    "\n",
    "## Import Packages\n",
    "\n",
    "TO DO: \n",
    "\n",
    "* Please reassign your corresponding paths\n",
    "    * Set path where data sets are stored; newly created data sets will also be stored there\n",
    "    * Set path to where the module is stored; this path will also be used to store the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/Bisa/Desktop/mlim_13/final_assignment\n"
     ]
    }
   ],
   "source": [
    "'TO DO: Please reassign your corresponding paths:'\n",
    "\n",
    "#set path where data sets are stored\n",
    "path_datasets = '/Users/Bisa/Desktop/mlim_13/final_assignment'\n",
    "#set path to where the module is stored\n",
    "%cd '/Users/Bisa/Desktop/mlim_13/final_assignment'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data wrangling \n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import time\n",
    "\n",
    "# feature engineering \n",
    "import module_p2v\n",
    "from module_clustering import p2cluster\n",
    "\n",
    "import module_negatives\n",
    "\n",
    "import module_lags\n",
    "\n",
    "import module_generate_dataset \n",
    "from module_generate_dataset import generate_dataset\n",
    "\n",
    "# train-test-splitting\n",
    "import module_train_test_splitting\n",
    "from module_train_test_splitting import train_test_splitting\n",
    "\n",
    "# heuristic model \n",
    "import module_baseline_heuristic_model\n",
    "# modeling \n",
    "import lightgbm as lgbm\n",
    "import xgboost as xgb\n",
    "import pickle\n",
    "\n",
    "import module_lightgbm\n",
    "from module_lightgbm import predict_lightgbm\n",
    "\n",
    "# generate data set for week 90\n",
    "import module_week90_generate_dataset\n",
    "from module_week90_generate_dataset import week90_generate_dataset\n",
    "\n",
    "# evaluation \n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import log_loss\n",
    "\n",
    "# visualiation\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# coupon recommendations \n",
    "import module_coupon_assignment \n",
    "from module_coupon_assignment import coupon_assignment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Set Creation and Feature Engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Basket Generation "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_baskets = pd.read_parquet(path_datasets + \"/baskets.parquet\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>products</th>\n",
       "      <th>price</th>\n",
       "      <th>basket_size</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[71, 91, 116, 123, 157, 167, 171, 184, 207, 225]</td>\n",
       "      <td>5908</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>[71, 91, 130, 157, 167, 191, 225]</td>\n",
       "      <td>4109</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>[52, 67, 71, 87, 91, 130, 184, 211, 238]</td>\n",
       "      <td>5108</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>[71, 91, 123, 130, 179, 184, 214, 225]</td>\n",
       "      <td>4681</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>[48, 67, 76, 91, 116, 123, 184, 225]</td>\n",
       "      <td>4740</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  week                                          products  price  \\\n",
       "0        0     0  [71, 91, 116, 123, 157, 167, 171, 184, 207, 225]   5908   \n",
       "1        0     1                 [71, 91, 130, 157, 167, 191, 225]   4109   \n",
       "2        0     2          [52, 67, 71, 87, 91, 130, 184, 211, 238]   5108   \n",
       "3        0     3            [71, 91, 123, 130, 179, 184, 214, 225]   4681   \n",
       "4        0     4              [48, 67, 76, 91, 116, 123, 184, 225]   4740   \n",
       "\n",
       "   basket_size  \n",
       "0           10  \n",
       "1            7  \n",
       "2            9  \n",
       "3            8  \n",
       "4            8  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate weekly item list per shopper \n",
    "def create_baskets(input_purchases):\n",
    "    \"\"\"Calculate an item list of products per shopper per week\"\"\"\n",
    "    input_origin = input_purchases.copy()\n",
    "    input_purchases = (\n",
    "        input_purchases.groupby(by=[\"shopper\", \"week\"])\n",
    "        .agg({\"product\": lambda x: list(x), \"price\": sum})\n",
    "        .reset_index(drop=False)\n",
    "    )\n",
    "    input_purchases = input_purchases.rename(columns={\"product\": \"products\"})\n",
    "    input_purchases[\"basket_size\"] = input_purchases[\"products\"].apply(len)\n",
    "    return input_purchases, input_origin\n",
    "\n",
    "baskets, _ = create_baskets(df_baskets)\n",
    "baskets.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Categories Generation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch #0\n",
      "Epoch #5\n",
      "Epoch #10\n",
      "Epoch #15\n",
      "Epoch #20\n",
      "Epoch #25\n",
      "Epoch #30\n",
      "Epoch #35\n",
      "Epoch #40\n",
      "Epoch #45\n",
      "Epoch #50\n",
      "Epoch #55\n",
      "Epoch #60\n",
      "Epoch #65\n",
      "Epoch #70\n",
      "Epoch #75\n",
      "Epoch #80\n",
      "Epoch #85\n",
      "Epoch #90\n",
      "Epoch #95\n"
     ]
    }
   ],
   "source": [
    "# train p2v model\n",
    "product_vector_model = module_p2v.p2v(baskets[\"products\"])\n",
    "product_vector_model.create_product_list()\n",
    "\n",
    "product_vector_model.train_p2v(vec_dim=30, epochs=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 430,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30-dimensional vector for 20:\n",
      "[ 16.91958     11.792615   -21.050528     7.352144     7.255684\n",
      "   1.0665644    9.210043    -4.9053097    5.138174    -1.5399554\n",
      " -15.364899    -2.210627     1.158818     4.2922053    5.387418\n",
      "   3.9973686   14.811209    -0.83350635   7.518798     5.9882135\n",
      " -11.342001     3.9756105  -16.87106      2.4546146   -3.8887858\n",
      " -10.50681      5.9898324    7.0195384   -4.0393276    7.58036   ]\n",
      "\n",
      "Most similar products for 20:\n",
      "[('29', 0.6650466322898865), ('23', 0.6482033729553223), ('22', 0.6268254518508911), ('27', 0.6112769246101379), ('25', 0.6083812713623047), ('21', 0.5770205855369568), ('26', 0.5235933065414429), ('28', 0.4557826519012451), ('24', 0.44312116503715515), ('232', 0.3516092002391815)]\n"
     ]
    }
   ],
   "source": [
    "product_vector_model.get_insights(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save backup\n",
    "# with open(\"product_vector_model.pickle\", \"wb\") as handle:\n",
    "#     pickle.dump(product_vector_model, handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"product_vector_model.pickle\", \"rb\") as handle:\n",
    "#     product_vector_model = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p2cluster_model = p2cluster(product_vector_model.p2v_model)\n",
    "p2cluster_model.tsne_train()\n",
    "p2cluster_model.tsne_plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p2cluster_model.elbow_plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# set to 25 clusters\n",
    "p2cluster_model.train_cluster(25)\n",
    "p2cluster_model.clust_plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "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>tsne_x</th>\n",
       "      <th>tsne_y</th>\n",
       "      <th>product</th>\n",
       "      <th>category_label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>3.079960</td>\n",
       "      <td>34.391747</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>38.935280</td>\n",
       "      <td>-93.043007</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>175</th>\n",
       "      <td>48.395184</td>\n",
       "      <td>50.941959</td>\n",
       "      <td>2</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206</th>\n",
       "      <td>-20.546894</td>\n",
       "      <td>-19.471403</td>\n",
       "      <td>3</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>-37.980141</td>\n",
       "      <td>3.057139</td>\n",
       "      <td>4</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        tsne_x     tsne_y product  category_label\n",
       "76    3.079960  34.391747       0              12\n",
       "65   38.935280 -93.043007       1               3\n",
       "175  48.395184  50.941959       2              22\n",
       "206 -20.546894 -19.471403       3              24\n",
       "37  -37.980141   3.057139       4              14"
      ]
     },
     "execution_count": 267,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "product_categories = p2cluster_model.get_categories()\n",
    "product_categories.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "metadata": {},
   "outputs": [],
   "source": [
    "product_categories[[\"product\", \"category_label\"]].to_csv(\n",
    "    \"product_categories.csv\", index=False\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preferences & Negative Samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"baskets_product_list.pickle\", \"rb\") as handle:\n",
    "#     baskets = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>products</th>\n",
       "      <th>price</th>\n",
       "      <th>basket_size</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[71, 91, 116, 123, 157, 167, 171, 184, 207, 225]</td>\n",
       "      <td>5908</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>[71, 91, 130, 157, 167, 191, 225]</td>\n",
       "      <td>4109</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>[52, 67, 71, 87, 91, 130, 184, 211, 238]</td>\n",
       "      <td>5108</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>[71, 91, 123, 130, 179, 184, 214, 225]</td>\n",
       "      <td>4681</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>[48, 67, 76, 91, 116, 123, 184, 225]</td>\n",
       "      <td>4740</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179961</th>\n",
       "      <td>1999</td>\n",
       "      <td>85</td>\n",
       "      <td>[21, 48, 86, 144, 158, 168, 192]</td>\n",
       "      <td>3674</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179962</th>\n",
       "      <td>1999</td>\n",
       "      <td>86</td>\n",
       "      <td>[4, 41, 64, 103, 132, 158, 220, 242]</td>\n",
       "      <td>4727</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179963</th>\n",
       "      <td>1999</td>\n",
       "      <td>87</td>\n",
       "      <td>[31, 47, 50, 64, 86, 103, 158, 168, 194]</td>\n",
       "      <td>5344</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179964</th>\n",
       "      <td>1999</td>\n",
       "      <td>88</td>\n",
       "      <td>[18, 47, 64, 144, 158, 168, 170, 192]</td>\n",
       "      <td>4712</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179965</th>\n",
       "      <td>1999</td>\n",
       "      <td>89</td>\n",
       "      <td>[4, 29, 31, 46, 75, 108, 114, 131, 144, 158, 1...</td>\n",
       "      <td>7515</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>179966 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        shopper  week                                           products  \\\n",
       "0             0     0   [71, 91, 116, 123, 157, 167, 171, 184, 207, 225]   \n",
       "1             0     1                  [71, 91, 130, 157, 167, 191, 225]   \n",
       "2             0     2           [52, 67, 71, 87, 91, 130, 184, 211, 238]   \n",
       "3             0     3             [71, 91, 123, 130, 179, 184, 214, 225]   \n",
       "4             0     4               [48, 67, 76, 91, 116, 123, 184, 225]   \n",
       "...         ...   ...                                                ...   \n",
       "179961     1999    85                   [21, 48, 86, 144, 158, 168, 192]   \n",
       "179962     1999    86               [4, 41, 64, 103, 132, 158, 220, 242]   \n",
       "179963     1999    87           [31, 47, 50, 64, 86, 103, 158, 168, 194]   \n",
       "179964     1999    88              [18, 47, 64, 144, 158, 168, 170, 192]   \n",
       "179965     1999    89  [4, 29, 31, 46, 75, 108, 114, 131, 144, 158, 1...   \n",
       "\n",
       "        price  basket_size  \n",
       "0        5908           10  \n",
       "1        4109            7  \n",
       "2        5108            9  \n",
       "3        4681            8  \n",
       "4        4740            8  \n",
       "...       ...          ...  \n",
       "179961   3674            7  \n",
       "179962   4727            8  \n",
       "179963   5344            9  \n",
       "179964   4712            8  \n",
       "179965   7515           13  \n",
       "\n",
       "[179966 rows x 5 columns]"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# set to 0..1999 shoppers, since the final prediction only assigns coupons to these shoppers\n",
    "baskets = baskets[baskets[\"shopper\"]<2000]\n",
    "baskets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0,  0,  0, ...,  0,  0, 13],\n",
       "       [ 6, 13,  0, ...,  0,  0,  6],\n",
       "       [ 0, 46,  0, ...,  0,  9,  5],\n",
       "       ...,\n",
       "       [ 0,  1,  0, ...,  0,  0,  6],\n",
       "       [ 0,  1,  1, ...,  1, 38,  0],\n",
       "       [ 0,  3,  0, ...,  0, 31,  0]], dtype=int8)"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate the total frequencies of product purchases per shopper per product\n",
    "negative_sampler = module_negatives.NegativeSampleGenerator(baskets,no_customers=2000)\n",
    "negative_sampler.calculate_frequencies()\n",
    "negative_sampler.get_total_frequency_without89()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\n",
      "[4, 11, 16, 38, 52, 56, 67, 71, 76, 81, 87, 91, 110, 116, 123, 130, 148, 157, 167, 171, 179, 181, 184, 191, 202, 207, 210, 211, 214, 225, 238, 244, 249]\n",
      "1:\n",
      "[0, 1, 4, 21, 22, 28, 33, 35, 36, 39, 42, 48, 50, 53, 63, 78, 83, 92, 94, 111, 116, 117, 125, 136, 142, 145, 146, 149, 156, 165, 168, 171, 182, 183, 189, 193, 199, 200, 209, 211, 212, 220, 228, 229, 235, 249]\n"
     ]
    }
   ],
   "source": [
    "# based on the frequencies: calculate the shopperXproduct combinations with a minimum purchase count of >=3\n",
    "preferences = negative_sampler.calculate_customer_preferences(min_frequency=3)\n",
    "negative_sampler.show_customer_preferences(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0% done.\n",
      "10% done.\n",
      "20% done.\n",
      "30% done.\n",
      "40% done.\n",
      "50% done.\n",
      "60% done.\n",
      "70% done.\n",
      "80% done.\n",
      "90% done.\n",
      "100% done.\n"
     ]
    }
   ],
   "source": [
    "# sample preferences products randomly in the size of the shoppers basket size of this week\n",
    "df_negative_samples = negative_sampler.generate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "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>week</th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>product_bought</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>238</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>130</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>181</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>52</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  week shopper product  product_bought\n",
       "0    0       0     238               0\n",
       "1    0       0      16               0\n",
       "2    0       0     130               0\n",
       "3    0       0     181               0\n",
       "4    0       0      52               0"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_negative_samples.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_negative_samples.to_parquet(\"df_negative_samples.parquet\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 315,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"preferences_min3.pickle\", \"wb\") as handle:\n",
    "#     pickle.dump(preferences, handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"df_negative_samples.pickle\", \"wb\") as handle:\n",
    "#     pickle.dump(df_negative_samples, handle)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Product Sampling (Week 90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"preferences_min3.pickle\", \"rb\") as handle:\n",
    "#     consumers_preferences = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "consumers_preferences = negative_sampler.cons_preferences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# sample all preferenced products with min_frequency >=3 per shopper for week 90\n",
    "def create_week_90_products():\n",
    "    df_90 = pd.DataFrame(columns=[\"week\", \"shopper\", \"product\", \"product_bought\"])\n",
    "    for customer in range(2000):\n",
    "        for pref in consumers_preferences[customer]:\n",
    "            df_90 = df_90.append({\"week\": 90 ,\n",
    "                                 \"shopper\": customer,\n",
    "                                 \"product\": pref,\n",
    "                                 \"product_bought\":0},ignore_index=True)\n",
    "    return df_90\n",
    "\n",
    "df_90 = create_week_90_products()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_90[\"shopper\"].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_90.to_parquet(\"df_90.parquet\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lagged Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_coupons = pd.read_parquet(path_datasets + \"/coupons.parquet\")\n",
    "df_full = pd.merge(df_baskets[df_baskets[\"shopper\"]<2000], df_coupons[df_coupons[\"shopper\"]<2000], on=['week', 'shopper', 'product'], how='outer')\n",
    "df_full['product_bought'] = np.where(df_full['price'].isna(), 0, 1)\n",
    "df_full = df_full[[\"week\",\"shopper\",\"product\",\"product_bought\"]]\n",
    "# df_90 = pd.read_parquet(\"df_90.parquet\") # only re-read if required\n",
    "df_full = pd.concat([df_full, df_90], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>week</th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>product_bought</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>52</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3665492</th>\n",
       "      <td>90</td>\n",
       "      <td>1999</td>\n",
       "      <td>228</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3665493</th>\n",
       "      <td>90</td>\n",
       "      <td>1999</td>\n",
       "      <td>237</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3665494</th>\n",
       "      <td>90</td>\n",
       "      <td>1999</td>\n",
       "      <td>242</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3665495</th>\n",
       "      <td>90</td>\n",
       "      <td>1999</td>\n",
       "      <td>245</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3665496</th>\n",
       "      <td>90</td>\n",
       "      <td>1999</td>\n",
       "      <td>249</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3665497 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         week  shopper  product  product_bought\n",
       "0           0        0        5               0\n",
       "1           0        0       16               0\n",
       "2           0        0       27               0\n",
       "3           0        0       35               0\n",
       "4           0        0       52               0\n",
       "...       ...      ...      ...             ...\n",
       "3665492    90     1999      228               0\n",
       "3665493    90     1999      237               0\n",
       "3665494    90     1999      242               0\n",
       "3665495    90     1999      245               0\n",
       "3665496    90     1999      249               0\n",
       "\n",
       "[3665497 rows x 4 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_full"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "purchase_lags = module_lags.LagCalculator(\n",
    "    df_full[[\"shopper\", \"product\", \"week\", \"product_bought\"]]\n",
    "    .sort_values(by=[\"shopper\", \"product\", \"week\"], ascending=[True, True, True])\n",
    "    .reset_index(drop=True)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0% done.\n",
      "10% done.\n",
      "20% done.\n",
      "30% done.\n",
      "40% done.\n",
      "50% done.\n",
      "60% done.\n",
      "70% done.\n",
      "80% done.\n",
      "90% done.\n"
     ]
    }
   ],
   "source": [
    "lags = purchase_lags.calculate_lags()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>week</th>\n",
       "      <th>product_bought</th>\n",
       "      <th>lag_weeks_of_product_per_customer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3001</th>\n",
       "      <td>1</td>\n",
       "      <td>146</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3002</th>\n",
       "      <td>1</td>\n",
       "      <td>146</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3003</th>\n",
       "      <td>1</td>\n",
       "      <td>146</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3004</th>\n",
       "      <td>1</td>\n",
       "      <td>146</td>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3005</th>\n",
       "      <td>1</td>\n",
       "      <td>146</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      shopper  product  week  product_bought  \\\n",
       "3001        1      146     5               0   \n",
       "3002        1      146     7               0   \n",
       "3003        1      146     8               0   \n",
       "3004        1      146    11               1   \n",
       "3005        1      146    12               0   \n",
       "\n",
       "      lag_weeks_of_product_per_customer  \n",
       "3001                                  4  \n",
       "3002                                  6  \n",
       "3003                                  7  \n",
       "3004                                 10  \n",
       "3005                                  1  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lags.iloc[3001:3006]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# replace \"product_bought\" of week 90 with NaN's\n",
    "lags.loc[lags['week']==90,['product_bought']] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>purchase_temporal_distribution</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>50.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>59.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>10.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>45.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>86.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  product  purchase_temporal_distribution\n",
       "0        0        4                       50.250000\n",
       "1        0       11                       59.666667\n",
       "2        0       15                       10.000000\n",
       "3        0       16                       45.000000\n",
       "4        0       21                       86.000000"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lags['week'] = lags['week'].astype('int32')\n",
    "purchase_temporal_distribution = purchase_lags.calculate_purchase_temporal_distribution(lags)\n",
    "purchase_temporal_distribution.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>avg_no_weeks_between_two_purchases</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>5.625000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>6.777778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>-1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>15.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>-1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  product  avg_no_weeks_between_two_purchases\n",
       "0        0        4                            5.625000\n",
       "1        0       11                            6.777778\n",
       "2        0       15                           -1.000000\n",
       "3        0       16                           15.500000\n",
       "4        0       21                           -1.000000"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_no_weeks_between_two_purchases = purchase_lags.calculate_avg_no_weeks_between_two_purchases(lags)\n",
    "avg_no_weeks_between_two_purchases.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# export\n",
    "lags.to_parquet(\"lags.parquet\")\n",
    "purchase_temporal_distribution.to_parquet(\"purchase_temporal_distribution.parquet\")\n",
    "avg_no_weeks_between_two_purchases.to_parquet(\"avg_no_weeks_between_two_purchases.parquet\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>week</th>\n",
       "      <th>product_bought</th>\n",
       "      <th>lag_weeks_of_product_per_customer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>606545</th>\n",
       "      <td>333</td>\n",
       "      <td>20</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>606738</th>\n",
       "      <td>333</td>\n",
       "      <td>41</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>606815</th>\n",
       "      <td>333</td>\n",
       "      <td>48</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>606836</th>\n",
       "      <td>333</td>\n",
       "      <td>51</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>606911</th>\n",
       "      <td>333</td>\n",
       "      <td>60</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607021</th>\n",
       "      <td>333</td>\n",
       "      <td>80</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607302</th>\n",
       "      <td>333</td>\n",
       "      <td>129</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607341</th>\n",
       "      <td>333</td>\n",
       "      <td>132</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607401</th>\n",
       "      <td>333</td>\n",
       "      <td>137</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607500</th>\n",
       "      <td>333</td>\n",
       "      <td>144</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607589</th>\n",
       "      <td>333</td>\n",
       "      <td>155</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607599</th>\n",
       "      <td>333</td>\n",
       "      <td>156</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607804</th>\n",
       "      <td>333</td>\n",
       "      <td>187</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>607881</th>\n",
       "      <td>333</td>\n",
       "      <td>192</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>608025</th>\n",
       "      <td>333</td>\n",
       "      <td>220</td>\n",
       "      <td>66</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>608115</th>\n",
       "      <td>333</td>\n",
       "      <td>227</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>608243</th>\n",
       "      <td>333</td>\n",
       "      <td>246</td>\n",
       "      <td>66</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        shopper  product  week  product_bought  \\\n",
       "606545      333       20    66             0.0   \n",
       "606738      333       41    66             0.0   \n",
       "606815      333       48    66             0.0   \n",
       "606836      333       51    66             0.0   \n",
       "606911      333       60    66             1.0   \n",
       "607021      333       80    66             0.0   \n",
       "607302      333      129    66             0.0   \n",
       "607341      333      132    66             0.0   \n",
       "607401      333      137    66             1.0   \n",
       "607500      333      144    66             1.0   \n",
       "607589      333      155    66             0.0   \n",
       "607599      333      156    66             0.0   \n",
       "607804      333      187    66             0.0   \n",
       "607881      333      192    66             1.0   \n",
       "608025      333      220    66             0.0   \n",
       "608115      333      227    66             1.0   \n",
       "608243      333      246    66             1.0   \n",
       "\n",
       "        lag_weeks_of_product_per_customer  \n",
       "606545                                 -1  \n",
       "606738                                  5  \n",
       "606815                                  1  \n",
       "606836                                 -1  \n",
       "606911                                  3  \n",
       "607021                                 -1  \n",
       "607302                                  8  \n",
       "607341                                  6  \n",
       "607401                                  3  \n",
       "607500                                 12  \n",
       "607589                                  5  \n",
       "607599                                 -1  \n",
       "607804                                 -1  \n",
       "607881                                  1  \n",
       "608025                                  5  \n",
       "608115                                 11  \n",
       "608243                                  2  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# example\n",
    "lags[(lags[\"shopper\"]==333) & (lags[\"week\"]==66)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Final Data Set Creation \n",
    "\n",
    "**In case the data sets are already loaded, proceed with code starting at 'Train-Test-Split'**\n",
    "\n",
    "* Prerequisites: the following data frames are needed. They shoud be named:\n",
    "    * baskets.parquet (provided/downloaded)\n",
    "    * coupons.parquet (provided/downloaded)\n",
    "    * df_negative_samples.parquet (created)\n",
    "    * product_categories.csv (created)\n",
    "    * avg_no_weeks_between_two_purchases.parquet (created)\n",
    "    * lags.parquet (created)\n",
    "    * purchase_temporal_distribution.parquet (created)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dataframes should be named: \n",
      "baskets.parquet, \n",
      "coupons.parquet, \n",
      "df_negative_samples.parquet, \n",
      "product_categories.csv, \n",
      "avg_no_weeks_between_two_purchases.parquet, \n",
      "lags.parquet and \n",
      "purchase_temporal_distribution.parquet\n",
      "\n",
      "Data sets (train and test) are generated and saved as a parquet file to: /Users/asmir/mlim_project/mlim/exercises/assignment\n"
     ]
    }
   ],
   "source": [
    "#the module outputs two dataframes: the train and test dataset\n",
    "#generate_dataset(path, train_start, train_end, test_start, test_end)\n",
    "\n",
    "train, test = generate_dataset(path_datasets, 0, 88, 1, 89)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>week</th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>price</th>\n",
       "      <th>discount</th>\n",
       "      <th>product_bought</th>\n",
       "      <th>category_label</th>\n",
       "      <th>avg_no_weeks_between_two_purchases</th>\n",
       "      <th>lag_weeks_of_product_per_customer</th>\n",
       "      <th>purchase_temporal_distribution</th>\n",
       "      <th>...</th>\n",
       "      <th>customer_prod_bought_dis_share</th>\n",
       "      <th>customer_prod_dis_offers</th>\n",
       "      <th>customer_prod_dis_offered_share</th>\n",
       "      <th>customer_product_share</th>\n",
       "      <th>customer_mean_product_price</th>\n",
       "      <th>customer_discount_buy_share</th>\n",
       "      <th>week_basket_size</th>\n",
       "      <th>week_basket_value</th>\n",
       "      <th>mean_basket_size</th>\n",
       "      <th>mean_basket_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>504.7</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>587.203947</td>\n",
       "      <td>0.030263</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5908.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5271.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>743.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>15.500000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.5</td>\n",
       "      <td>39</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.005263</td>\n",
       "      <td>587.203947</td>\n",
       "      <td>0.030263</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5908.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5271.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>27</td>\n",
       "      <td>407.4</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>587.203947</td>\n",
       "      <td>0.030263</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5908.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5271.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>391.3</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>587.203947</td>\n",
       "      <td>0.030263</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5908.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5271.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>52</td>\n",
       "      <td>607.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>9.111111</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>38.555556</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>38</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.011842</td>\n",
       "      <td>587.203947</td>\n",
       "      <td>0.030263</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5908.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5271.6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   week  shopper  product  price  discount  product_bought  category_label  \\\n",
       "0     0        0        5  504.7      30.0               0              12   \n",
       "1     0        0       16  743.0       0.0               0              13   \n",
       "2     0        0       27  407.4      30.0               0               4   \n",
       "3     0        0       35  391.3      35.0               0               9   \n",
       "4     0        0       52  607.0       0.0               0               7   \n",
       "\n",
       "   avg_no_weeks_between_two_purchases  lag_weeks_of_product_per_customer  \\\n",
       "0                           -1.000000                               -1.0   \n",
       "1                           15.500000                               -1.0   \n",
       "2                           -1.000000                               -1.0   \n",
       "3                           -1.000000                               -1.0   \n",
       "4                            9.111111                               -1.0   \n",
       "\n",
       "   purchase_temporal_distribution  ...  customer_prod_bought_dis_share  \\\n",
       "0                       -1.000000  ...                            -1.0   \n",
       "1                       45.000000  ...                             0.5   \n",
       "2                       -1.000000  ...                            -1.0   \n",
       "3                       -1.000000  ...                            -1.0   \n",
       "4                       38.555556  ...                             0.0   \n",
       "\n",
       "   customer_prod_dis_offers  customer_prod_dis_offered_share  \\\n",
       "0                         2                              0.0   \n",
       "1                        39                              0.5   \n",
       "2                         3                              0.0   \n",
       "3                         3                              0.0   \n",
       "4                        38                              0.0   \n",
       "\n",
       "   customer_product_share  customer_mean_product_price  \\\n",
       "0               -1.000000                   587.203947   \n",
       "1                0.005263                   587.203947   \n",
       "2               -1.000000                   587.203947   \n",
       "3               -1.000000                   587.203947   \n",
       "4                0.011842                   587.203947   \n",
       "\n",
       "   customer_discount_buy_share  week_basket_size  week_basket_value  \\\n",
       "0                     0.030263              10.0             5908.0   \n",
       "1                     0.030263              10.0             5908.0   \n",
       "2                     0.030263              10.0             5908.0   \n",
       "3                     0.030263              10.0             5908.0   \n",
       "4                     0.030263              10.0             5908.0   \n",
       "\n",
       "   mean_basket_size  mean_basket_value  \n",
       "0          8.978947             5271.6  \n",
       "1          8.978947             5271.6  \n",
       "2          8.978947             5271.6  \n",
       "3          8.978947             5271.6  \n",
       "4          8.978947             5271.6  \n",
       "\n",
       "[5 rows x 35 columns]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "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>week</th>\n",
       "      <th>shopper</th>\n",
       "      <th>product</th>\n",
       "      <th>price</th>\n",
       "      <th>discount</th>\n",
       "      <th>product_bought</th>\n",
       "      <th>category_label</th>\n",
       "      <th>avg_no_weeks_between_two_purchases</th>\n",
       "      <th>lag_weeks_of_product_per_customer</th>\n",
       "      <th>purchase_temporal_distribution</th>\n",
       "      <th>...</th>\n",
       "      <th>customer_prod_bought_dis_share</th>\n",
       "      <th>customer_prod_dis_offers</th>\n",
       "      <th>customer_prod_dis_offered_share</th>\n",
       "      <th>customer_product_share</th>\n",
       "      <th>customer_mean_product_price</th>\n",
       "      <th>customer_discount_buy_share</th>\n",
       "      <th>week_basket_size</th>\n",
       "      <th>week_basket_value</th>\n",
       "      <th>mean_basket_size</th>\n",
       "      <th>mean_basket_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>468.65</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>586.607895</td>\n",
       "      <td>0.032895</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4109.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5265.639474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>635.40</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>586.607895</td>\n",
       "      <td>0.032895</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4109.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5265.639474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>570.75</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>23.666667</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>39.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>41</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.003947</td>\n",
       "      <td>586.607895</td>\n",
       "      <td>0.032895</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4109.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5265.639474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>629.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>23</td>\n",
       "      <td>1.303030</td>\n",
       "      <td>1.0</td>\n",
       "      <td>43.030303</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030303</td>\n",
       "      <td>72</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.086842</td>\n",
       "      <td>586.607895</td>\n",
       "      <td>0.032895</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4109.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5265.639474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>81</td>\n",
       "      <td>513.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>6.666667</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>44.666667</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>29</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.003947</td>\n",
       "      <td>586.607895</td>\n",
       "      <td>0.032895</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4109.0</td>\n",
       "      <td>8.978947</td>\n",
       "      <td>5265.639474</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   week  shopper  product   price  discount  product_bought  category_label  \\\n",
       "0     1        0        5  468.65      35.0               0              12   \n",
       "1     1        0       34  635.40      10.0               0               9   \n",
       "2     1        0       56  570.75      25.0               0              14   \n",
       "3     1        0       71  629.00       0.0               1              23   \n",
       "4     1        0       81  513.00       0.0               0              21   \n",
       "\n",
       "   avg_no_weeks_between_two_purchases  lag_weeks_of_product_per_customer  \\\n",
       "0                           -1.000000                               -1.0   \n",
       "1                           -1.000000                               -1.0   \n",
       "2                           23.666667                               -1.0   \n",
       "3                            1.303030                                1.0   \n",
       "4                            6.666667                               -1.0   \n",
       "\n",
       "   purchase_temporal_distribution  ...  customer_prod_bought_dis_share  \\\n",
       "0                       -1.000000  ...                       -1.000000   \n",
       "1                       -1.000000  ...                       -1.000000   \n",
       "2                       39.000000  ...                        0.000000   \n",
       "3                       43.030303  ...                        0.030303   \n",
       "4                       44.666667  ...                        0.000000   \n",
       "\n",
       "   customer_prod_dis_offers  customer_prod_dis_offered_share  \\\n",
       "0                         1                              0.0   \n",
       "1                         5                              0.0   \n",
       "2                        41                              0.0   \n",
       "3                        72                              1.0   \n",
       "4                        29                              0.0   \n",
       "\n",
       "   customer_product_share  customer_mean_product_price  \\\n",
       "0               -1.000000                   586.607895   \n",
       "1               -1.000000                   586.607895   \n",
       "2                0.003947                   586.607895   \n",
       "3                0.086842                   586.607895   \n",
       "4                0.003947                   586.607895   \n",
       "\n",
       "   customer_discount_buy_share  week_basket_size  week_basket_value  \\\n",
       "0                     0.032895               7.0             4109.0   \n",
       "1                     0.032895               7.0             4109.0   \n",
       "2                     0.032895               7.0             4109.0   \n",
       "3                     0.032895               7.0             4109.0   \n",
       "4                     0.032895               7.0             4109.0   \n",
       "\n",
       "   mean_basket_size  mean_basket_value  \n",
       "0          8.978947        5265.639474  \n",
       "1          8.978947        5265.639474  \n",
       "2          8.978947        5265.639474  \n",
       "3          8.978947        5265.639474  \n",
       "4          8.978947        5265.639474  \n",
       "\n",
       "[5 rows x 35 columns]"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train-Test-Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The following features will be removed from the data sets (besides the target variable product_bought): \n",
      "shopper, \n",
      "product, \n",
      "purchase_w/o_dis, \n",
      "no_purchase_w_dis, \\discount_offered, \n",
      "discount_effect, \n",
      "week_basket_size and \n",
      "week_basket_value. \n",
      "Among others, reasons are target leakage and non-reproducibility for week 90.\n",
      "Training Observations: 3540798\n",
      "Testing Observations: 39535\n",
      "Observations: 3580333\n"
     ]
    }
   ],
   "source": [
    "#parameter are path, train_start, train_end, test_start, test_end, eval_set = False\n",
    "X_train, X_test, y_train, y_test = train_test_splitting(path_datasets, 0, 88, 89, 89, eval_set = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Heuristic Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'module_baseline_heuristic_model' from '/Users/asmir/mlim_project/mlim/exercises/assignment/module_baseline_heuristic_model.py'>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import importlib\n",
    "importlib.reload(module_baseline_heuristic_model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Assumptions**:\n",
    "1.  A customer will continue buying products with a coupon, if she has done so previously. \n",
    "\n",
    "**Logic/Algorithm**:\n",
    "\n",
    "1. Calculate the rate an item is bought with a coupon for each customer\n",
    "2. Order the items descending by coupon redemption rate\n",
    "3. Return the top 5 products\n",
    "4. Assign a coupon to these products for the week t+1, here week 90"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "baseline_data = module_baseline_heuristic_model.generate_heuristic_data(path_datasets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "heur_idx = module_baseline_heuristic_model.heuristic_model(path_datasets, baseline_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>coupon</th>\n",
       "      <th>product</th>\n",
       "      <th>discount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>185</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>2</td>\n",
       "      <td>207</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>3</td>\n",
       "      <td>179</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>4</td>\n",
       "      <td>214</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>75</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>157</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>2</td>\n",
       "      <td>165</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>3</td>\n",
       "      <td>220</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>4</td>\n",
       "      <td>53</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  week  coupon  product  discount\n",
       "0        0    90       0      185        30\n",
       "1        0    90       1       16        25\n",
       "2        0    90       2      207        20\n",
       "3        0    90       3      179        15\n",
       "4        0    90       4      214        15\n",
       "5        1    90       0       75        30\n",
       "6        1    90       1      157        25\n",
       "7        1    90       2      165        20\n",
       "8        1    90       3      220        15\n",
       "9        1    90       4       53        15"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "heur_idx.head(n=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\teval-auc:0.87180\ttrain-auc:0.87716\n",
      "[1]\teval-auc:0.87380\ttrain-auc:0.87867\n",
      "[2]\teval-auc:0.87658\ttrain-auc:0.88059\n",
      "[3]\teval-auc:0.87848\ttrain-auc:0.88105\n",
      "[4]\teval-auc:0.87713\ttrain-auc:0.88406\n",
      "[5]\teval-auc:0.87542\ttrain-auc:0.88546\n",
      "[6]\teval-auc:0.87281\ttrain-auc:0.88765\n",
      "[7]\teval-auc:0.87147\ttrain-auc:0.88849\n",
      "[8]\teval-auc:0.86856\ttrain-auc:0.88967\n",
      "[9]\teval-auc:0.86607\ttrain-auc:0.89156\n",
      "[10]\teval-auc:0.86506\ttrain-auc:0.89223\n",
      "[11]\teval-auc:0.86289\ttrain-auc:0.89261\n",
      "[12]\teval-auc:0.86272\ttrain-auc:0.89307\n",
      "[13]\teval-auc:0.86246\ttrain-auc:0.89348\n",
      "[14]\teval-auc:0.86219\ttrain-auc:0.89374\n",
      "[15]\teval-auc:0.86144\ttrain-auc:0.89454\n",
      "[16]\teval-auc:0.86174\ttrain-auc:0.89495\n",
      "[17]\teval-auc:0.86236\ttrain-auc:0.89516\n",
      "[18]\teval-auc:0.86191\ttrain-auc:0.89550\n",
      "[19]\teval-auc:0.86082\ttrain-auc:0.89586\n",
      "[20]\teval-auc:0.86009\ttrain-auc:0.89612\n",
      "[21]\teval-auc:0.85902\ttrain-auc:0.89642\n",
      "[22]\teval-auc:0.85965\ttrain-auc:0.89740\n",
      "[23]\teval-auc:0.85883\ttrain-auc:0.89794\n",
      "[24]\teval-auc:0.85996\ttrain-auc:0.89838\n",
      "[25]\teval-auc:0.86035\ttrain-auc:0.89872\n",
      "[26]\teval-auc:0.86075\ttrain-auc:0.89907\n",
      "[27]\teval-auc:0.85974\ttrain-auc:0.89933\n",
      "[28]\teval-auc:0.85908\ttrain-auc:0.89962\n",
      "[29]\teval-auc:0.85949\ttrain-auc:0.89964\n",
      "[30]\teval-auc:0.85980\ttrain-auc:0.89988\n",
      "[31]\teval-auc:0.85941\ttrain-auc:0.90009\n",
      "[32]\teval-auc:0.85916\ttrain-auc:0.90022\n",
      "[33]\teval-auc:0.85919\ttrain-auc:0.90050\n",
      "[34]\teval-auc:0.85846\ttrain-auc:0.90063\n",
      "[35]\teval-auc:0.85840\ttrain-auc:0.90084\n",
      "[36]\teval-auc:0.85860\ttrain-auc:0.90107\n",
      "[37]\teval-auc:0.85816\ttrain-auc:0.90118\n",
      "[38]\teval-auc:0.85868\ttrain-auc:0.90131\n",
      "[39]\teval-auc:0.85965\ttrain-auc:0.90165\n",
      "[40]\teval-auc:0.85979\ttrain-auc:0.90185\n",
      "[41]\teval-auc:0.85985\ttrain-auc:0.90202\n",
      "[42]\teval-auc:0.85978\ttrain-auc:0.90223\n",
      "[43]\teval-auc:0.86141\ttrain-auc:0.90260\n",
      "[44]\teval-auc:0.86146\ttrain-auc:0.90267\n",
      "[45]\teval-auc:0.86152\ttrain-auc:0.90279\n",
      "[46]\teval-auc:0.86157\ttrain-auc:0.90284\n",
      "[47]\teval-auc:0.86345\ttrain-auc:0.90303\n",
      "[48]\teval-auc:0.86327\ttrain-auc:0.90317\n",
      "[49]\teval-auc:0.86342\ttrain-auc:0.90329\n",
      "[50]\teval-auc:0.86449\ttrain-auc:0.90342\n",
      "[51]\teval-auc:0.86486\ttrain-auc:0.90355\n",
      "[52]\teval-auc:0.86551\ttrain-auc:0.90376\n",
      "[53]\teval-auc:0.86620\ttrain-auc:0.90392\n",
      "[54]\teval-auc:0.86626\ttrain-auc:0.90410\n",
      "[55]\teval-auc:0.86638\ttrain-auc:0.90423\n",
      "[56]\teval-auc:0.86725\ttrain-auc:0.90440\n",
      "[57]\teval-auc:0.86710\ttrain-auc:0.90456\n",
      "[58]\teval-auc:0.86713\ttrain-auc:0.90474\n",
      "[59]\teval-auc:0.86742\ttrain-auc:0.90489\n",
      "[60]\teval-auc:0.86748\ttrain-auc:0.90506\n",
      "[61]\teval-auc:0.86808\ttrain-auc:0.90515\n",
      "[62]\teval-auc:0.86816\ttrain-auc:0.90521\n",
      "[63]\teval-auc:0.86821\ttrain-auc:0.90527\n",
      "[64]\teval-auc:0.86919\ttrain-auc:0.90550\n",
      "[65]\teval-auc:0.86913\ttrain-auc:0.90566\n",
      "[66]\teval-auc:0.86918\ttrain-auc:0.90582\n",
      "[67]\teval-auc:0.87017\ttrain-auc:0.90614\n",
      "[68]\teval-auc:0.87024\ttrain-auc:0.90628\n",
      "[69]\teval-auc:0.87017\ttrain-auc:0.90638\n",
      "[70]\teval-auc:0.87023\ttrain-auc:0.90643\n",
      "[71]\teval-auc:0.86993\ttrain-auc:0.90654\n",
      "[72]\teval-auc:0.86995\ttrain-auc:0.90657\n",
      "[73]\teval-auc:0.87007\ttrain-auc:0.90662\n",
      "[74]\teval-auc:0.87025\ttrain-auc:0.90671\n",
      "[75]\teval-auc:0.86977\ttrain-auc:0.90678\n",
      "[76]\teval-auc:0.86986\ttrain-auc:0.90683\n",
      "[77]\teval-auc:0.86978\ttrain-auc:0.90695\n",
      "[78]\teval-auc:0.86981\ttrain-auc:0.90696\n",
      "[79]\teval-auc:0.86986\ttrain-auc:0.90699\n",
      "[80]\teval-auc:0.87017\ttrain-auc:0.90715\n",
      "[81]\teval-auc:0.87033\ttrain-auc:0.90725\n",
      "[82]\teval-auc:0.87039\ttrain-auc:0.90730\n",
      "[83]\teval-auc:0.87046\ttrain-auc:0.90741\n",
      "[84]\teval-auc:0.87035\ttrain-auc:0.90763\n",
      "[85]\teval-auc:0.87050\ttrain-auc:0.90789\n",
      "[86]\teval-auc:0.87057\ttrain-auc:0.90802\n",
      "[87]\teval-auc:0.87130\ttrain-auc:0.90811\n",
      "[88]\teval-auc:0.87140\ttrain-auc:0.90822\n",
      "[89]\teval-auc:0.87043\ttrain-auc:0.90836\n",
      "[90]\teval-auc:0.87054\ttrain-auc:0.90847\n",
      "[91]\teval-auc:0.87063\ttrain-auc:0.90857\n",
      "[92]\teval-auc:0.87089\ttrain-auc:0.90874\n",
      "[93]\teval-auc:0.87057\ttrain-auc:0.90895\n",
      "[94]\teval-auc:0.87063\ttrain-auc:0.90902\n",
      "[95]\teval-auc:0.87067\ttrain-auc:0.90913\n",
      "[96]\teval-auc:0.87078\ttrain-auc:0.90920\n",
      "[97]\teval-auc:0.87079\ttrain-auc:0.90929\n",
      "[98]\teval-auc:0.87075\ttrain-auc:0.90937\n",
      "[99]\teval-auc:0.87118\ttrain-auc:0.90942\n"
     ]
    }
   ],
   "source": [
    "xgb_mtx_train = xgb.DMatrix(X_train, label=y_train)\n",
    "xgb_mtx_test = xgb.DMatrix(X_test, label=y_test)\n",
    "params = {'max_depth': 10 , 'eta': 0.01, 'objective': 'binary:logistic', 'eval_metric': 'auc'}\n",
    "evallist = [(xgb_mtx_test, 'eval'), (xgb_mtx_train, 'train')]\n",
    "num_round = 100\n",
    "bst = xgb.train(params, xgb_mtx_train, num_round, evallist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix: \n",
      " [[21127  3215]\n",
      " [ 4546 10647]]\n"
     ]
    }
   ],
   "source": [
    "ypred = bst.predict(xgb_mtx_test)\n",
    "ytrue = y_test\n",
    "\n",
    "print(\"Confusion Matrix: \\n\", confusion_matrix(list(ytrue),list(ypred>0.5)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC:  0.871180772610282\n"
     ]
    }
   ],
   "source": [
    "#AUC\n",
    "print(\"AUC: \", metrics.auc(fpr, tpr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross Entropy:  0.6767952511705434\n"
     ]
    }
   ],
   "source": [
    "#Logloss\n",
    "print(\"Cross Entropy: \", log_loss(list(ytrue), list(ypred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training until validation scores don't improve for 50 rounds\n",
      "[200]\ttraining's binary_logloss: 0.411992\ttraining's auc: 0.891586\tvalid_1's binary_logloss: 0.391143\tvalid_1's auc: 0.908403\n",
      "Did not meet early stopping. Best iteration is:\n",
      "[300]\ttraining's binary_logloss: 0.392153\ttraining's auc: 0.895027\tvalid_1's binary_logloss: 0.358007\tvalid_1's auc: 0.91929\n",
      "defaultdict(<class 'collections.OrderedDict'>, {'training': OrderedDict([('binary_logloss', 0.3921530708235825), ('auc', 0.8950266281960872)]), 'valid_1': OrderedDict([('binary_logloss', 0.35800743808282914), ('auc', 0.9192897473995523)])})\n",
      "\n",
      "The computation took 9.42 minutes.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix: \n",
      " [[21101  3366]\n",
      " [ 3048 12279]]\n",
      "The AUC score is:  0.9192897473995523\n",
      "The Binary_Log_Loss is:  0.3580074380828356\n"
     ]
    }
   ],
   "source": [
    "#parameters are X_train, X_test, y_train, y_test, X_eval = None, y_eval = None, eval_set = False, output_probabilities = True, n_estimators = 300, early_stopping_rounds = 50, num_leaves = 1000, \n",
    "#reg_alpha = 0, reg_lambda = 0.5, subsample = 0.5, learning_rate = 0.01, verbose = 200). All parameter that is not specified here are the default parameters of the lightgbm model.\n",
    "\n",
    "lightgbm_model = predict_lightgbm(X_train, X_test, y_train, y_test, \n",
    "                         output_probabilities = True,\n",
    "                         n_estimators = 300, \n",
    "                         early_stopping_rounds = 50, \n",
    "                         num_leaves = 1000,\n",
    "                         reg_alpha = 0,\n",
    "                         reg_lambda = 0.5,\n",
    "                         subsample = 0.5, \n",
    "                         learning_rate = 0.01,\n",
    "                         verbose = 200)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification Accuracy:  0.7933982547110155\n",
      "Classification Error: \t 0.20660174528898445\n",
      "Recall:  \t\t 0.6850523267294149\n",
      "Precision: \t\t 0.7546950910013777\n",
      "F1 score: \t\t 0.7181893458459839\n",
      "AUC: \t\t\t 0.8846233294727821\n",
      "Cross Entropy: \t\t 0.41213648762970256\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Predictions and Classification Report\n",
    "lgbm_pred = lightgbm_model.predict(X_test)\n",
    "# Probabilities are returned for each class, unfortunately without the class names\n",
    "lgbm_pred_prob = lightgbm_model.predict_proba(X_test)\n",
    "lgbm_pred_prob = lgbm_pred_prob[:,1]\n",
    "\n",
    "print('Classification Accuracy: ',   metrics.accuracy_score(y_test, lgbm_pred))\n",
    "print('Classification Error: \\t',    1 - metrics.accuracy_score(y_test, lgbm_pred))\n",
    "print(\"Recall:  \\t\\t\",               metrics.recall_score(y_test, lgbm_pred))\n",
    "print(\"Precision: \\t\\t\",             metrics.precision_score(y_test, lgbm_pred))\n",
    "print(\"F1 score: \\t\\t\",              metrics.f1_score(y_test, lgbm_pred))\n",
    "print('AUC: \\t\\t\\t',                 metrics.roc_auc_score(y_test, lgbm_pred_prob))\n",
    "print('Cross Entropy: \\t\\t',       metrics.log_loss(y_test, lgbm_pred_prob))\n",
    "\n",
    "cm = confusion_matrix(y_test,lgbm_pred)\n",
    "ax = sns.heatmap(cm,cmap='viridis_r', annot=True, fmt='d', square=True)\n",
    "ax.set_xlabel('Predicted')\n",
    "ax.set_ylabel('True');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LightGBM vs Baseline Heuristic Model "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### LightGBM Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load LightGBM model \n",
    "filename = 'lightgbm_model_test89.pkl'\n",
    "with open(filename, 'rb') as file:\n",
    "    lgbm_m = pickle.load(file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Week 89 data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_89 = test[test['week'] == 89]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_89 = test_89.drop(['product_bought', 'shopper', 'product', 'purchase_w/o_dis', 'no_purchase_w_dis', 'discount_effect', 'week_basket_size', 'week_basket_value', 'discount_offered'], axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Light GBM - Coupon Recommendations for Week 89 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "importlib.reload(module_coupon_assignment)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating test data including prediction for various discounts d e {15,20,25,30}\n",
    "X_test_d15, X_test_d20, X_test_d25, X_test_d30 = module_coupon_assignment.get_test_discounts(X_test_89)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating predictions for each discount {15,20,25,30}\n",
    "pred_d15, pred_d20, pred_d25, pred_d30 = module_coupon_assignment.predictions(lgbm_m, X_test_d15, X_test_d20, X_test_d25, X_test_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge predictions back to original test data sets for discounts \n",
    "X_test_d15, X_test_d20, X_test_d25, X_test_d30 = module_coupon_assignment.merge_predictions(test_89, X_test_d15, X_test_d20, X_test_d25, X_test_d30, pred_d15, pred_d20, pred_d25, pred_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "# assign top 5 coupons via revenue maximization \n",
    "lgbm_coupons = module_coupon_assignment.coupon_assignment(X_test_d15, X_test_d20, X_test_d25, X_test_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>coupon</th>\n",
       "      <th>product</th>\n",
       "      <th>discount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>0</td>\n",
       "      <td>184</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>1</td>\n",
       "      <td>71</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>2</td>\n",
       "      <td>130</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>3</td>\n",
       "      <td>91</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>4</td>\n",
       "      <td>87</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>0</td>\n",
       "      <td>63</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>1</td>\n",
       "      <td>78</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>2</td>\n",
       "      <td>165</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>3</td>\n",
       "      <td>200</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>4</td>\n",
       "      <td>220</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  week  coupon  product  discount\n",
       "0        0    89       0      184        15\n",
       "1        0    89       1       71        15\n",
       "2        0    89       2      130        15\n",
       "3        0    89       3       91        15\n",
       "4        0    89       4       87        15\n",
       "5        1    89       0       63        15\n",
       "6        1    89       1       78        15\n",
       "7        1    89       2      165        15\n",
       "8        1    89       3      200        15\n",
       "9        1    89       4      220        25"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lgbm_coupons.head(n=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Baseline Heuristic Model - Coupons Recommendations for Week 89"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The baseline heuristic model applies the same logic irregardless of which week is being predicted, thus we can set week to an arbitrary value, here 89. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>coupon</th>\n",
       "      <th>product</th>\n",
       "      <th>discount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>0</td>\n",
       "      <td>185</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>2</td>\n",
       "      <td>207</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>3</td>\n",
       "      <td>179</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>4</td>\n",
       "      <td>214</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>0</td>\n",
       "      <td>75</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>1</td>\n",
       "      <td>157</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>2</td>\n",
       "      <td>165</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>3</td>\n",
       "      <td>220</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>4</td>\n",
       "      <td>53</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  week  coupon  product  discount\n",
       "0        0    89       0      185        30\n",
       "1        0    89       1       16        25\n",
       "2        0    89       2      207        20\n",
       "3        0    89       3      179        15\n",
       "4        0    89       4      214        15\n",
       "5        1    89       0       75        30\n",
       "6        1    89       1      157        25\n",
       "7        1    89       2      165        20\n",
       "8        1    89       3      220        15\n",
       "9        1    89       4       53        15"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "baseline_coupons = heur_idx.copy()\n",
    "baseline_coupons['week'] = 89\n",
    "baseline_coupons.head(n=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparison: LGBM vs Baseline Heuristic Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_89_s = test_89[['shopper', 'week', 'product', 'discount']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_89_s = test_89_s[test_89_s['discount'] > 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "shopper_prods_lgbm = pd.merge(lgbm_coupons, test_89_s, how='inner', on=(['shopper','product']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "shopper_prods_baseline = pd.merge(baseline_coupons, test_89_s, how='inner', on=(['shopper','product']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LGBM: Coupon Classification Accuracy:  0.06253126563281641\n"
     ]
    }
   ],
   "source": [
    "print('LGBM: Coupon Classification Accuracy: ',   (len(shopper_prods_lgbm)/len(test_89_s)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Baseline Heuristic Model: Coupon Classification Accuracy:  0.032416208104052026\n"
     ]
    }
   ],
   "source": [
    "print('Baseline Heuristic Model: Coupon Classification Accuracy: ',   (len(shopper_prods_baseline)/len(test_89_s)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The LGBM coupon solution is 93% better than the baseline solution.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Final Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = 'lightgbm_model_final.pkl'\n",
    "with open(filename, 'rb') as file:\n",
    "    model = pickle.load(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'coupon_assignment' from '/Users/Bisa/Documents/Studium/Masterstudium/5_Semester/MLiM/FinalAssignment/coupon_assignment.py'>"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import module_coupon_assignment \n",
    "#from module_coupon_assignment import coupon_assigmnet \n",
    "import module_week90_generate_dataset\n",
    "from module_week90_generate_dataset import week90_generate_dataset\n",
    "\n",
    "import importlib\n",
    "importlib.reload(module_coupon_assignment)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Trained LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = 'lightgbm_model_final.pkl'\n",
    "lightgbm_model = module_coupon_assignment.load_model(filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate Data Set for Week90 to Make Final Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dataframes should be named: \n",
      "baskets.parquet, \n",
      "coupons.parquet, \n",
      "df_negative_samples.parquet, \n",
      "product_categories.csv, \n",
      "avg_no_weeks_between_two_purchases.parquet, \n",
      "lags.parquet and \n",
      "purchase_temporal_distribution.parquet\n",
      "\n",
      "The data set for week 90 is generated and saved as a parquet file to: /Users/Bisa/Documents/Studium/Masterstudium/5_Semester/MLiM/datasets\n"
     ]
    }
   ],
   "source": [
    "# read week 90 code\n",
    "# for example \n",
    "X_test_90_all = week90_generate_dataset(path_datasets) # by funciton call - uses week90_gerate_dataset() \n",
    "X_test_90 = X_test_90_all.drop(['product_bought', 'shopper', 'product'], axis = 1) \n",
    "\n",
    "X_test_90 = pd.read_parquet(\"X_test_90.parquet\") # by reading csv \n",
    "\n",
    "#the variables price and discount will be added to the data set by the logical layer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Coupon Recommendation "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating test data including prediction for various discounts d e {15,20,25,30}\n",
    "X_test_d15, X_test_d20, X_test_d25, X_test_d30 = module_coupon_assignment.get_test_discounts(X_test_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating predictions for each discount {15,20,25,30}\n",
    "pred_d15, pred_d20, pred_d25, pred_d30 = module_coupon_assignment.predictions(lightgbm_model, X_test_d15, X_test_d20, X_test_d25, X_test_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge predictions back to original test data sets for discounts \n",
    "X_test_d15, X_test_d20, X_test_d25, X_test_d30 = module_coupon_assignment.merge_predictions(X_test_90_all, X_test_d15, X_test_d20, X_test_d25, X_test_d30, pred_d15, pred_d20, pred_d25, pred_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "# assign top 5 coupons via revenue maximization \n",
    "top5coupons = module_coupon_assignment.coupon_assignment(X_test_d15, X_test_d20, X_test_d25, X_test_d30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "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>shopper</th>\n",
       "      <th>week</th>\n",
       "      <th>coupon</th>\n",
       "      <th>product</th>\n",
       "      <th>discount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>2</td>\n",
       "      <td>116</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>3</td>\n",
       "      <td>38</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>4</td>\n",
       "      <td>244</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>117</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>63</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>2</td>\n",
       "      <td>116</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>3</td>\n",
       "      <td>111</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>4</td>\n",
       "      <td>36</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   shopper  week  coupon  product  discount\n",
       "0        0    90       0       56        15\n",
       "1        0    90       1       16        15\n",
       "2        0    90       2      116        15\n",
       "3        0    90       3       38        15\n",
       "4        0    90       4      244        15\n",
       "5        1    90       0      117        15\n",
       "6        1    90       1       63        15\n",
       "7        1    90       2      116        15\n",
       "8        1    90       3      111        15\n",
       "9        1    90       4       36        15"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top5coupons.head(n = 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "#save final prediction\n",
    "top5coupons.to_parquet(path_datasets + '/coupon_index.parquet')"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}