P5_algo_reduc_clustering-RFM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "79aa997d",
   "metadata": {},
   "source": [
    "## URL : \n",
    "\n",
    "    _  : https://www.definitions-marketing.com/definition/segmentation-rfm/\n",
    "    _  : https://www.kaggle.com/code/kautumn06/yellowbrick-clustering-evaluation-examples/notebook\n",
    "    _  : https://scikit-learn.org/stable/modules/generated/sklearn.metrics.adjusted_rand_score.html\n",
    "    _  : https://towardsdatascience.com/an-rfm-customer-segmentation-with-python-cf7be647733d\n",
    "    _  : https://towardsdatascience.com/recency-frequency-monetary-model-with-python-and-how-sephora-uses-it-to-optimize-their-google-d6a0707c5f17\n",
    "    - : https://machinelearningmastery.com/calculate-feature-importance-with-python/"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f4035dc4",
   "metadata": {},
   "source": [
    "## RFM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "97a16440",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import scipy.stats as st\n",
    "import warnings\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)\n",
    "pd.set_option(\"display.max_column\", 100)\n",
    "pd.set_option(\"display.max_rows\", 100)\n",
    "pd.set_option(\"display.max_colwidth\", None)\n",
    "import matplotlib.style\n",
    "plt.style.use('fast')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "781c6da7",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('clean_dataset.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bcec77dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# List of categorical features\n",
    "columns_to_convert_to_categorical = [\n",
    "    'customer_id',\n",
    "    'customer_unique_id',\n",
    "    'customer_city',\n",
    "    'customer_state',\n",
    "    'order_id',\n",
    "    'customer_id',\n",
    "    'order_status',\n",
    "    'order_id',\n",
    "    'product_id',\n",
    "    'seller_id',\n",
    "    'order_id',\n",
    "    'payment_type',\n",
    "    'review_id',\n",
    "    'order_id',\n",
    "    'product_id',\n",
    "    'product_category_name',\n",
    "    'seller_id',\n",
    "    'product_category_name'\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "89453d67",
   "metadata": {},
   "outputs": [],
   "source": [
    "# List of columns to convert to datetime\n",
    "columns_to_convert_to_datetime = [\n",
    "    'order_purchase_timestamp',\n",
    "    'order_approved_at',\n",
    "    'order_delivered_carrier_date',\n",
    "    'order_delivered_customer_date',\n",
    "    'order_estimated_delivery_date',\n",
    "    'shipping_limit_date',\n",
    "    'review_creation_date',\n",
    "    'review_answer_timestamp']\n",
    "\n",
    "# Converting dtype to datetime features\n",
    "for column_name in columns_to_convert_to_datetime:\n",
    "    data[column_name] = data[column_name].astype('datetime64[ns]')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "fdefc8bf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 119143 entries, 0 to 119142\n",
      "Data columns (total 45 columns):\n",
      " #   Column                         Non-Null Count   Dtype         \n",
      "---  ------                         --------------   -----         \n",
      " 0   Unnamed: 0                     119143 non-null  int64         \n",
      " 1   order_id                       119143 non-null  object        \n",
      " 2   customer_id                    119143 non-null  object        \n",
      " 3   order_status                   119143 non-null  object        \n",
      " 4   order_purchase_timestamp       119143 non-null  datetime64[ns]\n",
      " 5   order_approved_at              118966 non-null  datetime64[ns]\n",
      " 6   order_delivered_carrier_date   117057 non-null  datetime64[ns]\n",
      " 7   order_delivered_customer_date  115722 non-null  datetime64[ns]\n",
      " 8   order_estimated_delivery_date  119143 non-null  datetime64[ns]\n",
      " 9   review_id                      118146 non-null  object        \n",
      " 10  review_score                   118146 non-null  float64       \n",
      " 11  review_creation_date           118146 non-null  datetime64[ns]\n",
      " 12  review_answer_timestamp        118146 non-null  datetime64[ns]\n",
      " 13  payment_sequential             119140 non-null  float64       \n",
      " 14  payment_type                   119137 non-null  object        \n",
      " 15  payment_installments           119140 non-null  float64       \n",
      " 16  payment_value                  119140 non-null  float64       \n",
      " 17  customer_unique_id             119143 non-null  object        \n",
      " 18  customer_city                  119143 non-null  object        \n",
      " 19  customer_state                 119143 non-null  object        \n",
      " 20  geolocation_zip_code_prefix    118820 non-null  float64       \n",
      " 21  customer_lat                   118820 non-null  float64       \n",
      " 22  customer_lng                   118820 non-null  float64       \n",
      " 23  geolocation_city               118820 non-null  object        \n",
      " 24  geolocation_state              118820 non-null  object        \n",
      " 25  order_item_id                  118310 non-null  float64       \n",
      " 26  product_id                     118310 non-null  object        \n",
      " 27  seller_id                      118310 non-null  object        \n",
      " 28  shipping_limit_date            118310 non-null  datetime64[ns]\n",
      " 29  price                          118310 non-null  float64       \n",
      " 30  freight_value                  117920 non-null  float64       \n",
      " 31  product_name_lenght            116601 non-null  float64       \n",
      " 32  product_description_lenght     116601 non-null  float64       \n",
      " 33  product_photos_qty             116601 non-null  float64       \n",
      " 34  product_weight_g               118290 non-null  float64       \n",
      " 35  product_length_cm              118290 non-null  float64       \n",
      " 36  product_height_cm              118290 non-null  float64       \n",
      " 37  product_width_cm               118290 non-null  float64       \n",
      " 38  product_category_name          116576 non-null  object        \n",
      " 39  volume                         118290 non-null  float64       \n",
      " 40  delivery_time                  115707 non-null  float64       \n",
      " 41  estimated_delivery_time        118966 non-null  float64       \n",
      " 42  approval_time                  118966 non-null  float64       \n",
      " 43  purchase_year                  119143 non-null  int64         \n",
      " 44  purchase_month                 119143 non-null  int64         \n",
      "dtypes: datetime64[ns](8), float64(21), int64(3), object(13)\n",
      "memory usage: 40.9+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "fd2ce33f",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data.drop(columns=['Unnamed: 0'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a3f2b4fc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "order_id                         0.000000\n",
       "purchase_year                    0.000000\n",
       "customer_state                   0.000000\n",
       "customer_city                    0.000000\n",
       "customer_unique_id               0.000000\n",
       "order_estimated_delivery_date    0.000000\n",
       "purchase_month                   0.000000\n",
       "customer_id                      0.000000\n",
       "order_purchase_timestamp         0.000000\n",
       "order_status                     0.000000\n",
       "payment_value                    0.002518\n",
       "payment_sequential               0.002518\n",
       "payment_installments             0.002518\n",
       "payment_type                     0.005036\n",
       "approval_time                    0.148561\n",
       "estimated_delivery_time          0.148561\n",
       "order_approved_at                0.148561\n",
       "geolocation_city                 0.271103\n",
       "customer_lat                     0.271103\n",
       "customer_lng                     0.271103\n",
       "geolocation_state                0.271103\n",
       "geolocation_zip_code_prefix      0.271103\n",
       "shipping_limit_date              0.699160\n",
       "price                            0.699160\n",
       "seller_id                        0.699160\n",
       "product_id                       0.699160\n",
       "order_item_id                    0.699160\n",
       "volume                           0.715946\n",
       "product_width_cm                 0.715946\n",
       "product_height_cm                0.715946\n",
       "product_weight_g                 0.715946\n",
       "product_length_cm                0.715946\n",
       "review_creation_date             0.836810\n",
       "review_answer_timestamp          0.836810\n",
       "review_score                     0.836810\n",
       "review_id                        0.836810\n",
       "freight_value                    1.026498\n",
       "order_delivered_carrier_date     1.750837\n",
       "product_photos_qty               2.133571\n",
       "product_name_lenght              2.133571\n",
       "product_description_lenght       2.133571\n",
       "product_category_name            2.154554\n",
       "order_delivered_customer_date    2.871339\n",
       "delivery_time                    2.883929\n",
       "dtype: float64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_missing = data.isna().sum().sort_values()/len(data)*100\n",
    "df_missing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "4b915c7d",
   "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>order_id</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>order_status</th>\n",
       "      <th>order_purchase_timestamp</th>\n",
       "      <th>order_approved_at</th>\n",
       "      <th>order_delivered_carrier_date</th>\n",
       "      <th>order_delivered_customer_date</th>\n",
       "      <th>order_estimated_delivery_date</th>\n",
       "      <th>review_id</th>\n",
       "      <th>review_score</th>\n",
       "      <th>review_creation_date</th>\n",
       "      <th>review_answer_timestamp</th>\n",
       "      <th>payment_sequential</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>payment_installments</th>\n",
       "      <th>payment_value</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>customer_city</th>\n",
       "      <th>customer_state</th>\n",
       "      <th>geolocation_zip_code_prefix</th>\n",
       "      <th>customer_lat</th>\n",
       "      <th>customer_lng</th>\n",
       "      <th>geolocation_city</th>\n",
       "      <th>geolocation_state</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>shipping_limit_date</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>product_name_lenght</th>\n",
       "      <th>product_description_lenght</th>\n",
       "      <th>product_photos_qty</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_length_cm</th>\n",
       "      <th>product_height_cm</th>\n",
       "      <th>product_width_cm</th>\n",
       "      <th>product_category_name</th>\n",
       "      <th>volume</th>\n",
       "      <th>delivery_time</th>\n",
       "      <th>estimated_delivery_time</th>\n",
       "      <th>approval_time</th>\n",
       "      <th>purchase_year</th>\n",
       "      <th>purchase_month</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>9ef432eb6251297304e76186b10a928d</td>\n",
       "      <td>delivered</td>\n",
       "      <td>2017-10-02 10:56:33</td>\n",
       "      <td>2017-10-02 11:07:15</td>\n",
       "      <td>2017-10-04 19:55:00</td>\n",
       "      <td>2017-10-10 21:25:13</td>\n",
       "      <td>2017-10-18</td>\n",
       "      <td>a54f0611adc9ed256b57ede6b6eb5114</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2017-10-11</td>\n",
       "      <td>2017-10-12 03:43:48</td>\n",
       "      <td>1.0</td>\n",
       "      <td>credit_card</td>\n",
       "      <td>1.0</td>\n",
       "      <td>18.12</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>3149.0</td>\n",
       "      <td>-23.574809</td>\n",
       "      <td>-46.587471</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>1.0</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>2017-10-06 11:07:15</td>\n",
       "      <td>3.400864</td>\n",
       "      <td>2.165619</td>\n",
       "      <td>40.0</td>\n",
       "      <td>268.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>housewares</td>\n",
       "      <td>1976.0</td>\n",
       "      <td>8.429144</td>\n",
       "      <td>15.536632</td>\n",
       "      <td>0.007431</td>\n",
       "      <td>2017</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>9ef432eb6251297304e76186b10a928d</td>\n",
       "      <td>delivered</td>\n",
       "      <td>2017-10-02 10:56:33</td>\n",
       "      <td>2017-10-02 11:07:15</td>\n",
       "      <td>2017-10-04 19:55:00</td>\n",
       "      <td>2017-10-10 21:25:13</td>\n",
       "      <td>2017-10-18</td>\n",
       "      <td>a54f0611adc9ed256b57ede6b6eb5114</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2017-10-11</td>\n",
       "      <td>2017-10-12 03:43:48</td>\n",
       "      <td>3.0</td>\n",
       "      <td>voucher</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.00</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>3149.0</td>\n",
       "      <td>-23.574809</td>\n",
       "      <td>-46.587471</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>1.0</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>2017-10-06 11:07:15</td>\n",
       "      <td>3.400864</td>\n",
       "      <td>2.165619</td>\n",
       "      <td>40.0</td>\n",
       "      <td>268.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>housewares</td>\n",
       "      <td>1976.0</td>\n",
       "      <td>8.429144</td>\n",
       "      <td>15.536632</td>\n",
       "      <td>0.007431</td>\n",
       "      <td>2017</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>9ef432eb6251297304e76186b10a928d</td>\n",
       "      <td>delivered</td>\n",
       "      <td>2017-10-02 10:56:33</td>\n",
       "      <td>2017-10-02 11:07:15</td>\n",
       "      <td>2017-10-04 19:55:00</td>\n",
       "      <td>2017-10-10 21:25:13</td>\n",
       "      <td>2017-10-18</td>\n",
       "      <td>a54f0611adc9ed256b57ede6b6eb5114</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2017-10-11</td>\n",
       "      <td>2017-10-12 03:43:48</td>\n",
       "      <td>2.0</td>\n",
       "      <td>voucher</td>\n",
       "      <td>1.0</td>\n",
       "      <td>18.59</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>3149.0</td>\n",
       "      <td>-23.574809</td>\n",
       "      <td>-46.587471</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>1.0</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>2017-10-06 11:07:15</td>\n",
       "      <td>3.400864</td>\n",
       "      <td>2.165619</td>\n",
       "      <td>40.0</td>\n",
       "      <td>268.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>housewares</td>\n",
       "      <td>1976.0</td>\n",
       "      <td>8.429144</td>\n",
       "      <td>15.536632</td>\n",
       "      <td>0.007431</td>\n",
       "      <td>2017</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>b0830fb4747a6c6d20dea0b8c802d7ef</td>\n",
       "      <td>delivered</td>\n",
       "      <td>2018-07-24 20:41:37</td>\n",
       "      <td>2018-07-26 03:24:27</td>\n",
       "      <td>2018-07-26 14:31:00</td>\n",
       "      <td>2018-08-07 15:27:45</td>\n",
       "      <td>2018-08-13</td>\n",
       "      <td>8d5266042046a06655c8db133d120ba5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2018-08-08</td>\n",
       "      <td>2018-08-08 18:37:50</td>\n",
       "      <td>1.0</td>\n",
       "      <td>boleto</td>\n",
       "      <td>1.0</td>\n",
       "      <td>141.46</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>barreiras</td>\n",
       "      <td>BA</td>\n",
       "      <td>47813.0</td>\n",
       "      <td>-12.169860</td>\n",
       "      <td>-44.988369</td>\n",
       "      <td>barreiras</td>\n",
       "      <td>BA</td>\n",
       "      <td>1.0</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>289cdb325fb7e7f891c38608bf9e0962</td>\n",
       "      <td>2018-07-30 03:24:27</td>\n",
       "      <td>4.776599</td>\n",
       "      <td>3.125005</td>\n",
       "      <td>29.0</td>\n",
       "      <td>178.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>perfumery</td>\n",
       "      <td>4693.0</td>\n",
       "      <td>12.502292</td>\n",
       "      <td>17.858021</td>\n",
       "      <td>1.279745</td>\n",
       "      <td>2018</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>47770eb9100c2d0c44946d9cf07ec65d</td>\n",
       "      <td>41ce2a54c0b03bf3443c3d931a367089</td>\n",
       "      <td>delivered</td>\n",
       "      <td>2018-08-08 08:38:49</td>\n",
       "      <td>2018-08-08 08:55:23</td>\n",
       "      <td>2018-08-08 13:50:00</td>\n",
       "      <td>2018-08-17 18:06:29</td>\n",
       "      <td>2018-09-04</td>\n",
       "      <td>e73b67b67587f7644d5bd1a52deb1b01</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-08-18</td>\n",
       "      <td>2018-08-22 19:07:58</td>\n",
       "      <td>1.0</td>\n",
       "      <td>credit_card</td>\n",
       "      <td>3.0</td>\n",
       "      <td>179.12</td>\n",
       "      <td>3a653a41f6f9fc3d2a113cf8398680e8</td>\n",
       "      <td>vianopolis</td>\n",
       "      <td>GO</td>\n",
       "      <td>75265.0</td>\n",
       "      <td>-16.746337</td>\n",
       "      <td>-48.514624</td>\n",
       "      <td>vianopolis</td>\n",
       "      <td>GO</td>\n",
       "      <td>1.0</td>\n",
       "      <td>aa4383b373c6aca5d8797843e5594415</td>\n",
       "      <td>4869f7a5dfa277a7dca6462dcf3b52b2</td>\n",
       "      <td>2018-08-13 08:55:23</td>\n",
       "      <td>5.074549</td>\n",
       "      <td>2.955951</td>\n",
       "      <td>46.0</td>\n",
       "      <td>232.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>420.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>auto</td>\n",
       "      <td>9576.0</td>\n",
       "      <td>9.382708</td>\n",
       "      <td>26.628206</td>\n",
       "      <td>0.011505</td>\n",
       "      <td>2018</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                       customer_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  9ef432eb6251297304e76186b10a928d   \n",
       "1  e481f51cbdc54678b7cc49136f2d6af7  9ef432eb6251297304e76186b10a928d   \n",
       "2  e481f51cbdc54678b7cc49136f2d6af7  9ef432eb6251297304e76186b10a928d   \n",
       "3  53cdb2fc8bc7dce0b6741e2150273451  b0830fb4747a6c6d20dea0b8c802d7ef   \n",
       "4  47770eb9100c2d0c44946d9cf07ec65d  41ce2a54c0b03bf3443c3d931a367089   \n",
       "\n",
       "  order_status order_purchase_timestamp   order_approved_at  \\\n",
       "0    delivered      2017-10-02 10:56:33 2017-10-02 11:07:15   \n",
       "1    delivered      2017-10-02 10:56:33 2017-10-02 11:07:15   \n",
       "2    delivered      2017-10-02 10:56:33 2017-10-02 11:07:15   \n",
       "3    delivered      2018-07-24 20:41:37 2018-07-26 03:24:27   \n",
       "4    delivered      2018-08-08 08:38:49 2018-08-08 08:55:23   \n",
       "\n",
       "  order_delivered_carrier_date order_delivered_customer_date  \\\n",
       "0          2017-10-04 19:55:00           2017-10-10 21:25:13   \n",
       "1          2017-10-04 19:55:00           2017-10-10 21:25:13   \n",
       "2          2017-10-04 19:55:00           2017-10-10 21:25:13   \n",
       "3          2018-07-26 14:31:00           2018-08-07 15:27:45   \n",
       "4          2018-08-08 13:50:00           2018-08-17 18:06:29   \n",
       "\n",
       "  order_estimated_delivery_date                         review_id  \\\n",
       "0                    2017-10-18  a54f0611adc9ed256b57ede6b6eb5114   \n",
       "1                    2017-10-18  a54f0611adc9ed256b57ede6b6eb5114   \n",
       "2                    2017-10-18  a54f0611adc9ed256b57ede6b6eb5114   \n",
       "3                    2018-08-13  8d5266042046a06655c8db133d120ba5   \n",
       "4                    2018-09-04  e73b67b67587f7644d5bd1a52deb1b01   \n",
       "\n",
       "   review_score review_creation_date review_answer_timestamp  \\\n",
       "0           4.0           2017-10-11     2017-10-12 03:43:48   \n",
       "1           4.0           2017-10-11     2017-10-12 03:43:48   \n",
       "2           4.0           2017-10-11     2017-10-12 03:43:48   \n",
       "3           4.0           2018-08-08     2018-08-08 18:37:50   \n",
       "4           5.0           2018-08-18     2018-08-22 19:07:58   \n",
       "\n",
       "   payment_sequential payment_type  payment_installments  payment_value  \\\n",
       "0                 1.0  credit_card                   1.0          18.12   \n",
       "1                 3.0      voucher                   1.0           2.00   \n",
       "2                 2.0      voucher                   1.0          18.59   \n",
       "3                 1.0       boleto                   1.0         141.46   \n",
       "4                 1.0  credit_card                   3.0         179.12   \n",
       "\n",
       "                 customer_unique_id customer_city customer_state  \\\n",
       "0  7c396fd4830fd04220f754e42b4e5bff     sao paulo             SP   \n",
       "1  7c396fd4830fd04220f754e42b4e5bff     sao paulo             SP   \n",
       "2  7c396fd4830fd04220f754e42b4e5bff     sao paulo             SP   \n",
       "3  af07308b275d755c9edb36a90c618231     barreiras             BA   \n",
       "4  3a653a41f6f9fc3d2a113cf8398680e8    vianopolis             GO   \n",
       "\n",
       "   geolocation_zip_code_prefix  customer_lat  customer_lng geolocation_city  \\\n",
       "0                       3149.0    -23.574809    -46.587471        sao paulo   \n",
       "1                       3149.0    -23.574809    -46.587471        sao paulo   \n",
       "2                       3149.0    -23.574809    -46.587471        sao paulo   \n",
       "3                      47813.0    -12.169860    -44.988369        barreiras   \n",
       "4                      75265.0    -16.746337    -48.514624       vianopolis   \n",
       "\n",
       "  geolocation_state  order_item_id                        product_id  \\\n",
       "0                SP            1.0  87285b34884572647811a353c7ac498a   \n",
       "1                SP            1.0  87285b34884572647811a353c7ac498a   \n",
       "2                SP            1.0  87285b34884572647811a353c7ac498a   \n",
       "3                BA            1.0  595fac2a385ac33a80bd5114aec74eb8   \n",
       "4                GO            1.0  aa4383b373c6aca5d8797843e5594415   \n",
       "\n",
       "                          seller_id shipping_limit_date     price  \\\n",
       "0  3504c0cb71d7fa48d967e0e4c94d59d9 2017-10-06 11:07:15  3.400864   \n",
       "1  3504c0cb71d7fa48d967e0e4c94d59d9 2017-10-06 11:07:15  3.400864   \n",
       "2  3504c0cb71d7fa48d967e0e4c94d59d9 2017-10-06 11:07:15  3.400864   \n",
       "3  289cdb325fb7e7f891c38608bf9e0962 2018-07-30 03:24:27  4.776599   \n",
       "4  4869f7a5dfa277a7dca6462dcf3b52b2 2018-08-13 08:55:23  5.074549   \n",
       "\n",
       "   freight_value  product_name_lenght  product_description_lenght  \\\n",
       "0       2.165619                 40.0                       268.0   \n",
       "1       2.165619                 40.0                       268.0   \n",
       "2       2.165619                 40.0                       268.0   \n",
       "3       3.125005                 29.0                       178.0   \n",
       "4       2.955951                 46.0                       232.0   \n",
       "\n",
       "   product_photos_qty  product_weight_g  product_length_cm  product_height_cm  \\\n",
       "0                 4.0             500.0               19.0                8.0   \n",
       "1                 4.0             500.0               19.0                8.0   \n",
       "2                 4.0             500.0               19.0                8.0   \n",
       "3                 1.0             400.0               19.0               13.0   \n",
       "4                 1.0             420.0               24.0               19.0   \n",
       "\n",
       "   product_width_cm product_category_name  volume  delivery_time  \\\n",
       "0              13.0            housewares  1976.0       8.429144   \n",
       "1              13.0            housewares  1976.0       8.429144   \n",
       "2              13.0            housewares  1976.0       8.429144   \n",
       "3              19.0             perfumery  4693.0      12.502292   \n",
       "4              21.0                  auto  9576.0       9.382708   \n",
       "\n",
       "   estimated_delivery_time  approval_time  purchase_year  purchase_month  \n",
       "0                15.536632       0.007431           2017              10  \n",
       "1                15.536632       0.007431           2017              10  \n",
       "2                15.536632       0.007431           2017              10  \n",
       "3                17.858021       1.279745           2018               7  \n",
       "4                26.628206       0.011505           2018               8  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c548008a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#data.loc[:, ['customer_unique_id', \"order_id\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "135de3eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         e481f51cbdc54678b7cc49136f2d6af7\n",
       "1         e481f51cbdc54678b7cc49136f2d6af7\n",
       "2         e481f51cbdc54678b7cc49136f2d6af7\n",
       "3         53cdb2fc8bc7dce0b6741e2150273451\n",
       "4         47770eb9100c2d0c44946d9cf07ec65d\n",
       "                        ...               \n",
       "119138    63943bddc261676b46f01ca7ac2f7bd8\n",
       "119139    83c1379a015df1e13d02aae0204711ab\n",
       "119140    11c177c8e97725db2631073c19f07b62\n",
       "119141    11c177c8e97725db2631073c19f07b62\n",
       "119142    66dea50a8b16d9b4dee7af250b4be1a5\n",
       "Name: order_id, Length: 119143, dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['order_id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "380fe754",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         7c396fd4830fd04220f754e42b4e5bff\n",
       "1         7c396fd4830fd04220f754e42b4e5bff\n",
       "2         7c396fd4830fd04220f754e42b4e5bff\n",
       "3         af07308b275d755c9edb36a90c618231\n",
       "4         3a653a41f6f9fc3d2a113cf8398680e8\n",
       "                        ...               \n",
       "119138    da62f9e57a76d978d02ab5362c509660\n",
       "119139    737520a9aad80b3fbbdad19b66b37b30\n",
       "119140    5097a5312c8b157bb7be58ae360ef43c\n",
       "119141    5097a5312c8b157bb7be58ae360ef43c\n",
       "119142    60350aa974b26ff12caad89e55993bd6\n",
       "Name: customer_unique_id, Length: 119143, dtype: object"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[\"customer_unique_id\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "880934bb",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0          18.12\n",
       "1           2.00\n",
       "2          18.59\n",
       "3         141.46\n",
       "4         179.12\n",
       "           ...  \n",
       "119138    195.00\n",
       "119139    271.01\n",
       "119140    441.16\n",
       "119141    441.16\n",
       "119142     86.86\n",
       "Name: payment_value, Length: 119143, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['payment_value']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e539a572",
   "metadata": {},
   "outputs": [],
   "source": [
    "data.set_index('customer_unique_id', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "07ffa54d",
   "metadata": {},
   "outputs": [],
   "source": [
    "now = data['order_purchase_timestamp'].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "6ea14ed1",
   "metadata": {},
   "outputs": [],
   "source": [
    "rfm = data.groupby(['customer_unique_id']).agg({\n",
    "        'order_purchase_timestamp': lambda x: (now - x.max()).days,\n",
    "        'order_id':'count',\n",
    "        'payment_value':'sum'})\n",
    "    \n",
    "rfm.rename(columns={'order_purchase_timestamp': 'RecencyScore',\n",
    "                        'order_id': 'FrequencyScore',\n",
    "                        'payment_value': 'MonetaryScore'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "f5a57935",
   "metadata": {},
   "outputs": [],
   "source": [
    "rfm['FrequencyScore'] = rfm['FrequencyScore'][rfm['FrequencyScore'] > 0]\n",
    "rfm['FrequencyScore'] = np.log(rfm['FrequencyScore'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4e547222",
   "metadata": {},
   "outputs": [],
   "source": [
    "rfm['MonetaryScore'] = rfm['MonetaryScore'][rfm['MonetaryScore'] > 0]\n",
    "rfm['MonetaryScore'] = np.log(rfm['MonetaryScore'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "78272d86",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "0000366f3b9a7992bf8c76cfdf3221e2    4.955123\n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f    3.302849\n",
       "0000f46a3911fa3c0805444483337064    4.456902\n",
       "0000f6ccb0745a6a4b88665a16c9f078    3.775516\n",
       "0004aac84e0df4da2b147fca70cf8255    5.282645\n",
       "                                      ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684    8.327204\n",
       "fffea47cd6d3cc0a88bd621562a9d061    4.437698\n",
       "ffff371b4d645b6ecea244b27531430a    4.722598\n",
       "ffff5962728ec6157033ef9805bacc48    4.895524\n",
       "ffffd2657e2aad2907e67c3e9daecbeb    4.270536\n",
       "Name: MonetaryScore, Length: 96096, dtype: float64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfm['MonetaryScore']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bfe7c06a",
   "metadata": {},
   "source": [
    "Récupération de l'identifiant, de la date et de la valeur de chaque commande. Après jointure sur l'identifiant client on peut calculer les variables RFM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "56fd3677",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(1, figsize=(15, 4))\n",
    "\n",
    "plt.subplot(1, 3, 1)\n",
    "sns.distplot(rfm.RecencyScore, label='Recency Score',\n",
    "                 bins=int(1 + np.log2(len(rfm))))\n",
    "plt.ylabel(\"%\")\n",
    "\n",
    "plt.subplot(1, 3, 2)\n",
    "sns.distplot(rfm.FrequencyScore, label='Frequency Score',\n",
    "                 bins=int(1 + np.log2(len(rfm))))\n",
    "plt.ylabel(\"%\")\n",
    "\n",
    "plt.subplot(1, 3, 3)\n",
    "sns.distplot(rfm.MonetaryScore, label='Monetary Score',\n",
    "                 bins=int(1 + np.log2(len(rfm))))\n",
    "plt.ylabel(\"%\")\n",
    "plt.style.use('ggplot')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4f4e3a9e",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAIAAAEJCAYAAAAO34XRAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAs/klEQVR4nO3deXxU9b3/8fdMNoKQkAWIBAkFAYtWwY1F1gLXld5bwYgbBETL4qUVQWOr0urtba3iguLFhy2QoBW1uBSvFxVFVuFqEEtFEZAtkJANkgBJJjP5/v7IL3MzMAkJZGaSfF/Pf+DMOXPO55yZ+ZzvvOfMxGGMMQIAAAAAAFZwhroAAAAAAAAQPAQBAAAAAABYhCAAAAAAAACLEAQAAAAAAGARggAAAAAAACxCEAAAAAAAgEXCG3uHw4cPN2r5xMREFRQUNHYzAUEtzbcOiVrqQi1nrqNLly4hrCb4GtuHpebz2AUb+20X9ju06MXN57EIJBv2UbJjP23YR8mO/azZx8b0Ya4IAAAAAADAIgQBAAAAAABYhCAAAAAAAACLEAQAAAAAAGARggAAAAAAACxCEAAAAAAAgEUIAgAAAAAAsAhBAAAAAAAAFiEIAAAAAADAIgQBAAAAAABYhCAAAAAAAACLEAQAAAAAAGARggAAAAAAACxCEAAAAAAAgEUIAgAAAAAAsAhBAAAAAAAAFiEIAAAAAADAIgQBAAAAAABYhCAAAAAAAACLhIe6gJZuxYoVOnToUKPvFxERocrKyiavJz8/X5LUsWPHkNbRWMnJyfrFL34R6jKAVmnFihWKjo7WDTfcEOpSAACtzNmOhWs0duxal+TkZI0bN+6c1gHYhCDgHB06dEgHd+9SZ0dVo+7nClA95ab6Ig9XSVFI62iMI4YLU4BA2rJli5xOJ0EAAKDJne1YuEZjx67+MJYEGo8goAl0dlTpjsiKUJchSXrNFSVJzaaehqipGQAAAC3PuYyFm2LsylgSaDziMwAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALBIeyJWvWLFC0dHRuuGGGwK5GQDN3IoVKyRJ48aNC3EldqqoqAh1CQAAPzg/ojaeDwimgAYBhw4dUkRERCA3AaAFOHToUKhLsJoxJtQlAAD84PyI2ng+IJj4agAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsEh7qAgDYq7i4WH/+858lSffcc49KSkq0YMEC/fKXv1RycnKIq0MozJo1y/v/BQsWNOttnO16GnO/hi57puXqmz9nzhy5XC5FRkbq6aef9pn3wAMPqLKyUhEREZo/f7739scee0zHjh1TXFycfve739W5nvXr1+utt95SamqqhgwZIkl66623tH79eg0bNkzjx4/XRx99pPfff19jx47VmDFjVFxcrKVLl2ry5MmKiYnxmT61R9Set2vXLmVkZCgtLU1t27bVf/3Xf2nGjBlKSkryWZ8k/f3vf9fq1as1ZswYjR07VpJO225t9c3zx9/yjV1HqLSUOluamuM6fvx4/e1vf9O4ceO0YsUKeTwehYczHEe13bt3S/Lt2a1Nu3btdPz4cUlSYmKijh49Ko/Ho7i4OD3wwAMyxui5555TYWGhnE6nwsLC1LlzZ02bNk0lJSV6/vnnlZCQoIiICE2YMEHLly+Xy+VSUVGRpk6dqrfeekt5eXlKS0vT5ZdfruLiYr300ku68847ZYzx9rea/48bN05vvPGGHA6Hpk6desaenZ2dHZSxajB6MVcEAAiZVatWaf/+/dq/f79WrVqlzMxMlZeXKyMjI9SlAdZwuVw+/9ZWWVnp82+NY8eOSZKOHj1a73reeustSdKbb77pvW39+vWSpHXr1kmS3n//fUnSypUrJVX3hR9++EGrVq06bfrUHlF73quvvipJWrZsmZYsWSJjjBYvXnza+iRp9erVkqSPP/7Ye5u/5Royzx9/yzd2HaHSUupsaWqOa0ZGhn744QdlZmbqhx9+UFFRUahLA4KqJgSQpIKCAnk8HknV55NVq1Zp1apVKiwslCRVVVWpsrJS2dnZ3nNARUWFDh8+rP379ysjI0P79+9XTk6OKioqtHjxYuXl5UmqPhdI1a+9nTt3etdd099q/p+Zman9+/dr3759DerZwRqrBqMXEwQACIni4mJt2bLFO71p0ybl5uZKknJzc3Xo0KFQlYYQOfUTkEB8ItJU2zjb9TTmfg1d9kzL1Tc9Z84cn3m1px944AGfeTXTjz32mM/t8+bN87uemjf8NTZs2OANBmo888wzPtMrV67U//7v/8oYoy1btujQoUPe6c2bN/v0iJ07d3rnbdy40TuY9Hg8KisrkySVlZXp888/966vpKREf//730/bZnFxsc92S0pKvPPrm+ePv+Ubu45QaSl1tjS1j2tubq7Pv6WlpXK73aEuEc1Aa74KoKE2bNigTZs2+Z1Xe5xY49Tpmt4vVZ8LNmzY4HMO2bJly2n/r72OzZs319uzs7OzgzJWDVYvDui1SPn5+aqsrAzY5Z2NFRERcdqnGucqOztbEcbRpOu0zVHjUGV2tp544okmf3zOViCeK2erNdSSnZ2tqKgon9tWrVrlHbhL1alvbRkZGfr1r399doUCaJBTrwKoPX3qa71muuZqgBq1rwqovZ5T3/TXviqgxr59+3ymP/74Y4WFhUmq7gkZGRne3nDqm6XFixd75xljTlt3jZplqqqqtGrVKm3YsOG0bZ48efK05VJTUyVV96q65vnjb3ljTKPWESqN3Vc0TO3jeipjjA4ePNhsxsoNcepYoDmMhWvGkk15HJvT+MsmdfXzul5D9XnzzTe95xS32y2Ho/p56vF4/G7H4/HU27MzMzN9lg/UWDVYvfiMVwSsXr1a6enpSk9Pb/KNA7BXVlZWvYP3U1Nem9GHYZPan+zn5ub6BIa1lZWV1TmvrvV++eWXfudlZWX5bLf2cvXNa+i6GruOUGkpdYbK2fbi2sfVn8Y8jwE0Tu3XV824s67xpzGm3p59pisSmkqwevEZrwgYPXq0Ro8efVYr79ixoyIiIjR9+vSzun9TS0xMVEFBQZOuc8GCBXLt2dmk67RNnMMosmtXPfroo03++JytQDxXzlZrqMVfQn/FFVdo06ZNdTbjpKSkRm+ntTqXPgy0NGFhYfJ4PAoLC1PHjh2Vn5/v941SdHS0XC5Xg99EhYWF6corrzztigCpuh9t3rzZu90rr7yyQfP88bd8zaWoDV1HqDR2X21ztr249nH1JyYmpkVdFn7qWKA5jIVrxpJNeRyDPf5qSc+BlqTmnCJJDodDxhjvv6dyOBz19uykpCSfN/+BGqsGqxfzGwEAQuK6667zXq4lSU6nbzuaNGlSsEsCrBMZGVnndEREhM+8mukOHTr43B4XF+d3PbfccovPbampqRo6dKjPbd27d/eZHjNmjLcXOJ1OTZo0yTt96i+rT5kyxTuv5nJPf2qv77rrrjvtjdyYMWN03XXXnbZcjfrm+eNv+cauI1RaSp0tTe3jeiqHw6H4+PggVwQ0X3X187peQ/VJTU31OYfUjDvDwsL8/rWOsLCwenv2xIkTfZYP1Fg1WL2YIABASMTGxmrAgAHe6cGDB3uT1aSkJP58oIVOvXIkEN+ZbaptnO16GnO/hi57puXqmz71zwXWnq795wJrTz/++OM+t//ud7/zu55T3/QPGTLktHBg9uzZPtNjx47V1VdfLYfDoQEDBig5Odk7PXDgQJ8e0adPH++8a665xmeAFx0dLan6qoFBgwZ51xcTE6Of/exnp20zNjbWZ7u1/1RTffP88bd8Y9cRKi2lzpam9nFNSkry+bd9+/b8+UBICtyfzG1JhgwZosGDB/udV3ucWOPU6ZreL1WfC4YMGeJzDhkwYMBp/6+9joEDB9bbs7t27RqUsWqwejFBAICQue6665SSkqKUlBRdd911mjhxotq0acPVAEAQ1Xyaf+qn+tL/XQVw6tUBNVcFxMXF1buemjf+tX/kqCYgGDZsmCTppptuklT9hlyq7gs9evTwfgJSe/rUHlF73p133ilJuuuuuzR58mQ5HA5NmTLltPVJ8l4VMGbMGO9t/pZryDx//C3f2HWESkups6WpOa6TJk1Sjx49NHHiRPXo0YOrAWCddu3aef+fmJjoDXHj4uK8n8YnJCRIqv40PCIiQl27dvWeA6KiotSlSxelpKRo0qRJSklJ0fnnn6+oqChNmTJFnTp1klR9LpCqX3t9+vTxrrumv9X8f+LEiUpJSVH37t0b1LODNVYNRi92mPp+rcuPw4cPN3jZBQsWWPMbAXdEVjTpes/Wa67qX2ZvLvU0xGuuKEX27KPHH3+8xX8XPhBaQy01KXdTff/t1Dq6dOnSJOttKRrTh6X/O+42ftrQnF4/wcR+26W57De9uPGPRVOfH4Ohrt8IONuxZ1OMXWvGki35NwKk4D8fmkvvCDQb9rNmHxvTh7kiAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYJD+TKk5OTFR0dHchNAGgBkpOTQ12C1RwOR6hLAAD4wfkRtfF8QDAFNAgYN26cEhMTVVBQEMjNAGjmxo0bF+oSrBYVFSWnkwvAAKC54fyI2ng+IJgYGQIAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIsQBAAAAAAAYBGCAAAAAAAALEIQAAAAAACARQgCAAAAAACwCEEAAAAAAAAWIQgAAAAAAMAiBAEAAAAAAFiEIAAAAAAAAIuEh7qA1uCIceo1V1Soy5BUXYukZlNPQxwxTl0Q6iIAAABwVs5lLNwUY1fGkkDjEQSco+Tk5LO6X0REhCorK5u4GqlNfr4kKbJjx5DW0RgX6OyPI4AzGzBggKKjo0NdBgCgFTrXMVxjx67+MJYEGo8g4ByNGzfurO6XmJiogoKCJq6m5dYBIHDGjRvHax0AEBBnOxYGEFr8RgAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAixAEAAAAAABgEYIAAAAAAAAsQhAAAAAAAIBFCAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAiziMMSbURQAAAAAAgOAI+BUB6enpgd5Eg1HL6ZpLHRK11IVaTtdc6mhJbD1m7Ldd2G+Emg2PhQ37KNmxnzbso2THfp7NPvLVAAAAAAAALEIQAAAAAACARQIeBIwePTrQm2gwajldc6lDopa6UMvpmksdLYmtx4z9tgv7jVCz4bGwYR8lO/bThn2U7NjPs9lHfiwQAAAAAACL8NUAAAAAAAAsQhAAAAAAAIBFwgO58m3btmnJkiWqqqrSqFGj9G//9m8B29ZLL72krVu3KjY2VvPnz5ckHT9+XM8++6zy8/PVsWNH3X///WrXrp0k6Z133tGnn34qp9OpyZMnq1+/fk1WS0FBgRYuXKhjx47J4XBo9OjRuuGGG0JSj8vl0rx58+R2u+XxeDRw4EClpqaG7NhUVVUpPT1d8fHxSk9PD1kdM2fOVJs2beR0OhUWFqY//vGPIavlxIkTWrRokQ4ePCiHw6Hp06erS5cuQa/l8OHDevbZZ73TeXl5Sk1N1fDhw4Ney/vvv69PP/1UDodDF1xwgWbMmCGXyxWSx6elC2Yfbk78nRNsUNf5p7Wr61xni1PPrQgum153NjzX/I3LevfuHeqympy/sVZkZGSoyzonjX0/2FL5289ly5YpKytL4eHh6ty5s2bMmKHzzjuv/hWZAPF4POa+++4zubm5prKy0syZM8ccPHgwUJsz33zzjdmzZ4+ZPXu297Zly5aZd955xxhjzDvvvGOWLVtmjDHm4MGDZs6cOcblcpkjR46Y++67z3g8niarpaioyOzZs8cYY8zJkyfNrFmzzMGDB0NST1VVlSkrKzPGGFNZWWkefvhhs3PnzpAdm5UrV5rnnnvO/OEPfzDGhO4xmjFjhikuLva5LVS1vPDCC2b16tXGmOrH6Pjx4yGrpYbH4zFTp041eXl5Qa+lsLDQzJgxw1RUVBhjjJk/f75Zs2ZNyI9JSxTsPtyc+Dsn2KCu809rV9e5zhannlsRXDa97mx4rvkbl7U2dY21WrrGvB9syfzt57Zt24zb7TbGVO9zQ/YzYF8N2L17t5KSktS5c2eFh4dr8ODB+uKLLwK1OfXt2/e0dOeLL77Q8OHDJUnDhw/3bv+LL77Q4MGDFRERoU6dOikpKUm7d+9uslri4uLUo0cPSVJ0dLSSk5NVVFQUknocDofatGkjSfJ4PPJ4PHI4HCGppbCwUFu3btWoUaO8t4XqMfInFLWcPHlS3377rX76059KksLDw3XeeeeF/Lhs375dSUlJ6tixY0hqqaqqksvlksfjkcvlUlxcXMiPSUsU7D7cnPg7J9igrvNPa1fXuc4G/s6tCC5bXnc2PNfqGpe1Rv7GWi1dY94PtmT+9vOyyy5TWFiYJKl3794N6kEB+2pAUVGREhISvNMJCQnatWtXoDbnV3FxsfdJHRcXp5KSEm9tvXr18i4XHx8fsIadl5envXv36sILLwxZPVVVVXrooYeUm5ura6+9Vr169QpJLUuXLtWdd96psrIy722hfIx+//vfS5LGjBmj0aNHh6SWvLw8xcTE6KWXXtL+/fvVo0cPpaWlhfy5u3HjRl1zzTWSgv8YxcfHa+zYsZo+fboiIyN12WWX6bLLLgv5MWmJmkMfRujUPv/YwN+5zgb+zq0Indb8urPhuVbXuKwmaGwt6hprtUZ1jR9bs08//VSDBw8+43IBuyLA+PmrhM0lnfdXWyCUl5dr/vz5SktLU9u2bUNWj9Pp1FNPPaVFixZpz549OnDgQNBrycrKUmxsrDcxP5NAH5MnnnhCTz75pH7961/rww8/1I4dO0JSi8fj0d69e/Uv//Iv+tOf/qSoqCi9++67IamlhtvtVlZWlgYOHFjvcoGq5fjx4/riiy+0cOFCvfzyyyovL9e6deuCXkdr0Jz7MAKroeef1qQx57rWorHnVgRWa37d2fJca+y4rKVq7FgLLcfbb7+tsLAwDR069IzLBuyKgISEBBUWFnqnCwsLg37JSWxsrI4ePaq4uDgdPXpUMTExfmsrKipSfHx8k27b7XZr/vz5Gjp0qAYMGBDyeiTpvPPOU9++fbVt27ag17Jz5059+eWX+uqrr+RyuVRWVqYFCxaE7JjUrCs2NlZXXXWVdu/eHZJaEhISlJCQ4P3kauDAgXr33XdD+lz56quv9KMf/UgdOnSQFPzn7fbt29WpUyfvdgYMGKDvv/8+5K+flqg59GEEn7/zj01qn+u6desW6nICqq5z66xZs0JdmnVa++vOludaXeOy1qausdawYcNCXFnTq2v82Bp99tlnysrK0mOPPdagD34CdkVAz549lZOTo7y8PLndbm3atElXXnlloDbn15VXXqm1a9dKktauXaurrrrKe/umTZtUWVmpvLw85eTkNOklXMYYLVq0SMnJybrppptCWk9JSYlOnDghqfpXlbdv367k5OSg13L77bdr0aJFWrhwoX71q1/pkksu0axZs0JyTMrLy72XtZWXl+sf//iHunXrFpJaOnTooISEBB0+fFhSdWPu2rVryJ67ku/XAmq2GcxaEhMTtWvXLlVUVMgYE7LnbGvQHPowgquu809rV9e5rrWr69yK4LLhdWfLc62ucVlrU9dYqzWqa/zY2mzbtk3vvfeeHnroIUVFRTXoPg4TwOtqt27dqoyMDFVVVWnkyJG6+eabA7UpPffcc9qxY4dKS0sVGxur1NRUXXXVVXr22WdVUFCgxMREzZ492/vDCm+//bbWrFkjp9OptLQ09e/fv8lq+e677/TYY4+pW7du3jTmtttuU69evYJez/79+7Vw4UJVVVXJGKNBgwZp/PjxKi0tDcmxkaRvvvlGK1euVHp6ekjqOHLkiJ5++mlJ1ZeADRkyRDfffHPIjsm+ffu0aNEiud1uderUSTNmzJAxJiS1VFRUaPr06XrxxRe9lzWG4ri8+eab2rRpk8LCwtS9e3dNmzZN5eXlIXvOtmTB7MPNib9zQs2PP7VmdZ1/Lr/88hBXFlh1netsUvvciuCy7XXX2p9r/sZlrfHHZ/2NtSIiIkJd1jlp7PvBlsrffr7zzjtyu93efevVq5fuvffeetcT0CAAAAAAAAA0LwH7agAAAAAAAGh+CAIAAAAAALAIQQAAAAAAABYhCAAAAAAAwCIEAQAAAAAAWIQgAAAAAAAAi4SHugAExsyZM3Xs2DE5nU61adNG/fr109133602bdqEurQGKyws1JIlS/Ttt9/K7XYrMTFRY8eO1YgRI0JdGgB41e63NZ5//nnFx8eHsKrA+uKLL/Tmm28qLy9P4eHhSklJ0bRp09SpU6dQlwbAcjNnzlRRUZFefvllxcTEeG+fO3eu9u/frxdffDFgveqzzz7TJ598oieeeCIg65fov2g6BAGt2EMPPaRLL71Ux44d0+9//3u98847uu2220JdVoO9+OKLSklJ0cKFCxUREaEDBw7o2LFjTboNj8ejsLCwJl0nAPvU9Ft/Wlufyc3N1Ysvvqg5c+bokksuUXl5ub7++mufIORcGWNkjGnSdQKwR6dOnbRx40Zdf/31kqQDBw7I5XKFuKozO9P5gv6LpkQQYIEOHTrosssu0759+yRJ33//vTIzM5Wdna2OHTsqLS1NF198sSTp+PHjyszM1Ndffy2Xy6Uf//jHevDBByVJWVlZWr58ufLz89W1a1fdc889SklJkVSdvl577bVat26d8vPz1a9fP82cOVORkZGSfNPLmJgY3X333SorK9O7776rJ5980lvrypUr9e233+rBBx/U7t27NWnSJO9VDD/60Y989uu7777Tq6++quzsbEVHR+vWW2/ViBEjdPLkSS1evFhfffWVoqKiNGrUKP385z+X0+n0JrU9e/bU2rVrde2112rcuHF6/fXX9fnnn8vtduuqq65SWlqat3YAaKzU1FRNmTJFH3zwgTwejxYuXFhvD927d68WLVqknJwc9e/fXw6HQ0lJSZowYYLfT5hSU1O1YMECJSUlqbKyss4e9s033+iFF17QjTfeqPfee09Op1O33XabRo4cKUlyuVxavny5Nm/erBMnTqhbt2569NFHNX/+fPXr1887iJakOXPmKDU1VVVVVerUqZN+8pOfSJKio6M1cOBA73JVVVV69913tWbNGhUXF+v888/X3LlzlZiYqJ07d2rp0qU6fPiwunTporS0NPXp00eS9Nvf/lZ9+vTRjh079MMPP2j+/PnyeDxavHixfvjhB8XExOjWW2/V4MGDA/74AWjZhg0bpnXr1nl72Geffabhw4dr+fLlktSgsWKvXr20Zs0atW3bVlOnTlX//v29983IyNBXX30lh8OhkSNHKjU1VYcPH9Yrr7wit9utu+66S2FhYVq6dKm2bt2q5cuX68iRI2rbtq13eUnKy8vTfffdp2nTpumtt95Sp06dvFfy0n8RcAat0owZM8zXX39tjDGmoKDAzJ492yxevNgUFhaayZMnm6ysLOPxeMzXX39tJk+ebIqLi40xxvznf/6neeaZZ0xpaamprKw033zzjTHGmD179pi7777bfP/998bj8Zg1a9aYGTNmGJfL5d1eenq6KSwsNKWlpeZXv/qV+fDDD40xxuzatctMnDjRfP3118bj8ZjCwkKTnZ1tXC6XSUtLMwcPHvTWPXfuXPP5558bY4x5/PHHzSOPPGI2bNhg8vPzffYvPz/f3HXXXWb9+vWmsrLSlJSUmL179xpjjHnhhRfMk08+aU6ePGmOHDliZs2aZT755BNjjDFr1qwxt956q/nggw+M2+02FRUVZsmSJeaPf/yjKS0tNSdPnjR/+MMfzGuvvRagRwZAa1O739a45ZZbzOOPP25KS0tNRUVFvT20srLSTJ8+3axcudJUVlaazz//3EyYMMG8/vrrxpjqvvXII4+ctv6cnBxjjKm3h/3zn/80t956q1m+fLmprKw0WVlZ5o477jClpaXGGGNeeeUVM2/ePFNYWGg8Ho/57rvvjMvlMhs3bjQPP/ywd3t79+41kydPNpWVlSY3N9fcfvvtZsmSJWb79u2mrKzMp7b33nvPzJ492xw6dMhUVVWZvXv3mpKSElNaWmrS0tLM2rVrjdvtNuvXrzdpaWmmpKTEGGPMvHnzzLRp08yBAweM2+02J06cMNOmTTOffvqpcbvdZs+ePWbKlCnmwIEDTfjoAWhtanryrFmzzMGDB43H4zHTpk0zeXl55pZbbjFHjhw541hxwoQJ5uOPPzYej8d8+OGH5t577zVVVVXGGGOefPJJ8/LLL5uysjJz7Ngxk56ebj766CPvfU/t1//85z/N/v37jcfjMfv27TNTp041W7ZsMcYYc+TIEXPLLbeYF154wZSVlZmKigr6L4KGaz5asaeeekoTJ07U9OnTFRsbq9TUVK1bt079+/fX5ZdfLqfTqUsvvVQ9e/bU1q1bdfToUW3btk333HOP2rVrp/DwcPXt21eS9Mknn2j06NHq1auXnE6nRowYofDwcO3atcu7veuvv17x8fFq166drrjiCu8VCJ9++qlGjhypSy+9VE6nU/Hx8UpOTlZERIQGDx6s9evXS5IOHjyo/Px8XXHFFZKk+++/XxdddJFWrFihmTNnau7cudq9e7ckaf369frJT36iIUOGKDw8XO3bt1f37t1VVVWlTZs26fbbb1d0dLQ6deqkm266SevWrfPWGRcXp+uvv15hYWGKiIjQJ598okmTJqldu3aKjo7WzTffrI0bNwbjIQLQSjz11FNKS0tTWlqa/vSnP0mSfv7zn6tdu3aKjIyst4d+//338ng8uvHGGxUeHq6BAweqZ8+eDdquMeaMPSwsLEzjx49XeHi4Lr/8crVp00aHDx9WVVWV1qxZo7S0NMXHx8vpdKpPnz6KiIjQVVddpZycHOXk5EiS1q1bp8GDBys8PFydO3fWvHnzVFRUpGeffVZ33323Fi5cqPLycknV54sJEyaoS5cucjgc6t69u9q3b6+tW7cqKSlJw4YNU1hYmIYMGaIuXbooKyvLW+uIESN0wQUXKCwsTNu2bVPHjh01cuRIhYWFqUePHhowYIA2b97cVA8bgFas5qqAf/zjH+rSpYv3d1saMlZMTEzU6NGj5XQ6NXz4cB09elTFxcU6duyYtm3bprS0NLVp00axsbG68cYbtWnTpjrruPjii9WtWzc5nU6lpKTommuu0Y4dO3yWueWWW9SmTRtFRkbSfxE0fDWgFZs7d64uvfRS7dixQ88//7xKS0tVUFCgzZs3+7zwPR6PLr74YhUWFqpdu3Zq167daesqKCjQ2rVrtWrVKu9tbrdbRUVF3ukOHTp4/x8ZGemdV1hY6L2c6lTDhw/X888/rwkTJmjdunUaNGiQIiIiJEnt2rXTHXfcoTvuuEMlJSVatmyZnnrqKS1atEiFhYXq3LnzaesrKSnx/rBgjY4dO/rUWXteSUmJKioqlJ6e7r3NGKOqqiq/9QKAPzX9tkZqaqoSEhK80/X1UIfDofj4eDkcDu+82n2qPg3pYe3bt/f5zmlUVJTKy8tVWlqqyspKJSUlnbbeiIgIDRo0SOvXr9f48eO1ceNGPfDAA975vXv31uzZsyVJu3fv1nPPPae3335bt99+e539uaioSB07dvS57dT+XPuY5efna9euXUpLS/Pe5vF4NGzYsIYcGgCWGzZsmObNm6e8vDwNHz7ce3tpaekZx4q1x7RRUVGSpPLych0/flwej0f33nuvd74xxqd3nWrXrl3661//qgMHDsjtdsvtdvtczi/59j76L4KFIMACffv21YgRI5SZmalevXpp6NChmjZt2mnLHT16VMePH9eJEyd03nnn+cxLSEjQzTffrJtvvrnR209ISFBubq7feb1791Z4eLi+/fZbbdiwQb/85S/9LhcTE6OxY8dq7dq1On78uBISErxXB5y6XFhYmAoKCtS1a1dJ1QPwun69u3379oqMjNQzzzzTqn/hG0Dw1X5jX18P3bFjh4qKimSM8d6nsLDQ+wY9KirK50euav9o6rn0sPbt2ysiIkK5ubnq3r37afNHjBihF154QRdddJGioqLUu3dvv+u58MILdfXVV+vAgQPefT1y5Ii6devms1x8fLy2bNnic1tBQYH69evnnT71mPXt21ePPvpoo/YLAKTqN7qdOnXSV1995TPurQlHGzpWrC0hIUHh4eH6y1/+0uAfgV2wYIGuvfZaPfzww4qMjNTSpUtVUlLis0zt3ifRfxEcfDXAEjfeeKO2b9+uiy66SFlZWdq2bZuqqqrkcrn0zTffqLCwUHFxcerXr5/+/Oc/6/jx43K73d5Ll0aNGqWPP/5Yu3btkjFG5eXl2rp1q8rKys647Z/+9Kf67LPPtH37dlVVVamoqEiHDh3yzh8+fLgWL16ssLAwXXTRRd7bX331VR04cEAej0dlZWX66KOPlJSUpPbt22vo0KHavn27Nm3aJI/Ho9LSUu3bt09Op1ODBg3S66+/rrKyMuXn5+v999/X0KFD/dbmdDo1atQoLV26VMXFxZKqU9Nt27adw9EGAF/19dDevXvL6XTqf/7nf+TxeLRlyxafoDMlJUUHDx7Uvn375HK59Oabb3rnnUsPczqdGjlypDIzM1VUVKSqqip9//33qqyslCRvXZmZmT6fAn333XdavXq1d3uHDh1SVlaWd6A6atQovfHGG8rJyZExRvv371dpaan69++vnJwcbdiwQR6PR5s2bVJ2drYuv/xyv/VdccUVysnJ0bp167yfou3evVvZ2dmNO/gArDVt2jQ99thjPn8+u7Fjxdri4uJ02WWXKTMzUydPnlRVVZVyc3O94+UOHTqoqKhIbrfbe5+ysjLv18R2796tDRs2nHE79F8EA1cEWCImJkbDhg3Tf//3f+vBBx/Uq6++queff15Op1MXXnih7rnnHknSv//7v2vp0qW6//775Xa7dfHFF6tv377q2bOnfvGLX2jx4sXKyclRZGSkLrroIv34xz8+47YvvPBCzZgxQxkZGcrLy1NsbKzuvvtuJScnS6q+dOuNN97QuHHjfO7ncrn09NNP6+jRo4qMjFSvXr28f8EgMTFRDz/8sJYtW6aXX35Zbdu21a233qru3btrypQpWrx4se677z5FRkZq1KhR3l/I9ueOO+7Q3/72N/3mN79RaWmp4uPjNWbMGJ+UFADORX09NDw8XHPmzNHLL7+s5cuXq3///rr66qu99+3SpYvGjx+vJ554QpGRkbrtttu0evVq7/xz6WETJ07UX//6Vz388MMqLy9X9+7d9Zvf/MY7v6Y/z50713tb27Zt9eWXX+qNN95QeXm5YmJiNGjQIP3sZz+TJN10002qrKzUf/zHf6i0tFTJycmaM2eOEhISlJ6eriVLluiVV15RUlKS0tPTff7Od23R0dF65JFHlJGRoYyMDBljlJKSokmTJjX28AOwlL+vPklq9Fixtvvuu0+vvfaaZs+erbKyMnXu3Fn/+q//Kkm65JJLvH8Vxul06i9/+YumTp2qzMxMLV68WH379tWgQYN04sSJM26H/otAcxhjTKiLgN1cLpemTp2qJ598Uueff36oywGAkFu4cKESEhI0YcKEkNaxdu1arV692udPFwIAAo/+i0DjqwEIuY8++kg9e/YkBACAZqSiokIfffSRRo8eHepSAMAq9F8EA0EAQmrmzJn64IMPNHHixFCXAgD4/7Zt26apU6cqNjZWQ4YMCXU5AGAN+i+Cha8GAAAAAABgEa4IAAAAAADAIgQBAAAAAABYhCAAAAAAAACLEAQAAAAAAGARggAAAAAAACzy/wDsRwBFxBwplQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1296x288 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(1, figsize=(18, 4))\n",
    "\n",
    "plt.subplot(131)\n",
    "sns.boxplot(rfm[\"RecencyScore\"])\n",
    "\n",
    "plt.subplot(132)\n",
    "sns.boxplot(rfm[\"FrequencyScore\"])\n",
    "\n",
    "plt.subplot(133)\n",
    "sns.boxplot(rfm[\"MonetaryScore\"])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6d0dedb4",
   "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>RecencyScore</th>\n",
       "      <th>FrequencyScore</th>\n",
       "      <th>MonetaryScore</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>96096.000000</td>\n",
       "      <td>96096.000000</td>\n",
       "      <td>96093.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>287.735691</td>\n",
       "      <td>0.131395</td>\n",
       "      <td>4.805777</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>153.414676</td>\n",
       "      <td>0.339076</td>\n",
       "      <td>0.918514</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.260721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>163.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.158883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>268.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.730392</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>397.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.315322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>772.000000</td>\n",
       "      <td>4.317488</td>\n",
       "      <td>11.601967</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       RecencyScore  FrequencyScore  MonetaryScore\n",
       "count  96096.000000    96096.000000   96093.000000\n",
       "mean     287.735691        0.131395       4.805777\n",
       "std      153.414676        0.339076       0.918514\n",
       "min        0.000000        0.000000       2.260721\n",
       "25%      163.000000        0.000000       4.158883\n",
       "50%      268.000000        0.000000       4.730392\n",
       "75%      397.000000        0.000000       5.315322\n",
       "max      772.000000        4.317488      11.601967"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfm.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "cee66dc9",
   "metadata": {
    "scrolled": true
   },
   "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>RecencyScore</th>\n",
       "      <th>FrequencyScore</th>\n",
       "      <th>MonetaryScore</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0000366f3b9a7992bf8c76cfdf3221e2</th>\n",
       "      <td>160</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.955123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000b849f77a49e4a4ce2b2a4ca5be3f</th>\n",
       "      <td>163</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.302849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f46a3911fa3c0805444483337064</th>\n",
       "      <td>585</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.456902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f6ccb0745a6a4b88665a16c9f078</th>\n",
       "      <td>369</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.775516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0004aac84e0df4da2b147fca70cf8255</th>\n",
       "      <td>336</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.282645</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  RecencyScore  FrequencyScore  MonetaryScore\n",
       "customer_unique_id                                                           \n",
       "0000366f3b9a7992bf8c76cfdf3221e2           160             0.0       4.955123\n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f           163             0.0       3.302849\n",
       "0000f46a3911fa3c0805444483337064           585             0.0       4.456902\n",
       "0000f6ccb0745a6a4b88665a16c9f078           369             0.0       3.775516\n",
       "0004aac84e0df4da2b147fca70cf8255           336             0.0       5.282645"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfm.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "7b9dc44a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 96096 entries, 0000366f3b9a7992bf8c76cfdf3221e2 to ffffd2657e2aad2907e67c3e9daecbeb\n",
      "Data columns (total 3 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   RecencyScore    96096 non-null  int64  \n",
      " 1   FrequencyScore  96096 non-null  float64\n",
      " 2   MonetaryScore   96093 non-null  float64\n",
      "dtypes: float64(2), int64(1)\n",
      "memory usage: 4.9+ MB\n"
     ]
    }
   ],
   "source": [
    "rfm.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f58dfb0a",
   "metadata": {},
   "source": [
    "##  Features clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83dfcdcc",
   "metadata": {},
   "source": [
    "Le datset est aggrégé par clients unique. Les variables catégorielles ne peuvent pas être aggrégés, on va les supprimer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "6510f1fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create shipping_time variable\n",
    "data['shipping_time'] = data['shipping_limit_date'] \\\n",
    "    - data['order_delivered_carrier_date']\n",
    "data['shipping_time'] = data['shipping_time'] \\\n",
    "    .dt.total_seconds() / 86400\n",
    "\n",
    "# Create review_delay variable \n",
    "data['review_delay'] = data['review_creation_date'] \\\n",
    "    - data['order_purchase_timestamp']\n",
    "data['review_delay'] = data['review_delay'] \\\n",
    "    .dt.total_seconds() / 86400\n",
    "\n",
    "# Create review_answer_delay variable \n",
    "data['review_answer_delay'] = data['review_answer_timestamp'] \\\n",
    "    - data['review_creation_date']\n",
    "data['review_answer_delay'] = data['review_answer_delay'] \\\n",
    "    .dt.total_seconds() / 86400"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "33cb9c6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Features to aggregate\n",
    "features = ['review_creation_date',\n",
    "            'review_answer_timestamp',\n",
    "            'estimated_delivery_time',\n",
    "            'delivery_time',\n",
    "            'approval_time',\n",
    "            'purchase_year',\n",
    "            'purchase_month',\n",
    "            'order_purchase_timestamp',\n",
    "            'payment_sequential',\n",
    "            'freight_value',\n",
    "            'price',\n",
    "            'payment_value',\n",
    "            'order_item_id',\n",
    "            'product_weight_g',\n",
    "            'product_photos_qty',\n",
    "            'product_name_lenght',\n",
    "            'volume',\n",
    "            'shipping_limit_date',\n",
    "            'order_delivered_carrier_date',\n",
    "            'shipping_time',\n",
    "            'review_delay',\n",
    "            'review_answer_delay'\n",
    "           ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "43974c69",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = data.groupby('customer_unique_id').agg({'review_creation_date': \"mean\",\n",
    "                                                         'review_answer_timestamp': \"mean\",\n",
    "                                                         'estimated_delivery_time': \"mean\",\n",
    "                                                         'delivery_time': 'mean',\n",
    "                                                         'approval_time': 'mean',\n",
    "                                                         'purchase_year': 'mean',\n",
    "                                                         'purchase_month': 'mean',\n",
    "                                                         'order_purchase_timestamp': 'mean',\n",
    "                                                         'payment_sequential': \"sum\",\n",
    "                                                         'freight_value': \"sum\",\n",
    "                                                         'price': 'sum',\n",
    "                                                         'payment_value': \"sum\",\n",
    "                                                         'order_item_id': 'sum',\n",
    "                                                         'product_weight_g': \"mean\",\n",
    "                                                         'product_photos_qty': 'mean',\n",
    "                                                         'product_name_lenght': 'mean',\n",
    "                                                         'volume': \"mean\",\n",
    "                                                         'shipping_limit_date': \"mean\",\n",
    "                                                         'order_delivered_carrier_date': \"mean\",\n",
    "                                                         'shipping_time':\"mean\",\n",
    "                                                         'review_delay':\"mean\",\n",
    "                                                         'review_answer_delay':\"mean\",\n",
    "                                                        })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4422cdbb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 96096 entries, 0000366f3b9a7992bf8c76cfdf3221e2 to ffffd2657e2aad2907e67c3e9daecbeb\n",
      "Data columns (total 22 columns):\n",
      " #   Column                        Non-Null Count  Dtype         \n",
      "---  ------                        --------------  -----         \n",
      " 0   review_creation_date          95380 non-null  datetime64[ns]\n",
      " 1   review_answer_timestamp       95380 non-null  datetime64[ns]\n",
      " 2   estimated_delivery_time       95997 non-null  float64       \n",
      " 3   delivery_time                 93343 non-null  float64       \n",
      " 4   approval_time                 95997 non-null  float64       \n",
      " 5   purchase_year                 96096 non-null  float64       \n",
      " 6   purchase_month                96096 non-null  float64       \n",
      " 7   order_purchase_timestamp      96096 non-null  datetime64[ns]\n",
      " 8   payment_sequential            96096 non-null  float64       \n",
      " 9   freight_value                 96096 non-null  float64       \n",
      " 10  price                         96096 non-null  float64       \n",
      " 11  payment_value                 96096 non-null  float64       \n",
      " 12  order_item_id                 96096 non-null  float64       \n",
      " 13  product_weight_g              95407 non-null  float64       \n",
      " 14  product_photos_qty            94108 non-null  float64       \n",
      " 15  product_name_lenght           94108 non-null  float64       \n",
      " 16  volume                        95407 non-null  float64       \n",
      " 17  shipping_limit_date           95420 non-null  datetime64[ns]\n",
      " 18  order_delivered_carrier_date  94466 non-null  datetime64[ns]\n",
      " 19  shipping_time                 94465 non-null  float64       \n",
      " 20  review_delay                  95380 non-null  float64       \n",
      " 21  review_answer_delay           95380 non-null  float64       \n",
      "dtypes: datetime64[ns](5), float64(17)\n",
      "memory usage: 16.9+ MB\n"
     ]
    }
   ],
   "source": [
    "model.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ccb5850b",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.columns = features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "fffcccb5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#model = model.join(rfm[['RecencyScore','FrequencyScore','MonetaryScore']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "39c30070",
   "metadata": {
    "scrolled": false
   },
   "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>review_creation_date</th>\n",
       "      <th>review_answer_timestamp</th>\n",
       "      <th>estimated_delivery_time</th>\n",
       "      <th>delivery_time</th>\n",
       "      <th>approval_time</th>\n",
       "      <th>purchase_year</th>\n",
       "      <th>purchase_month</th>\n",
       "      <th>order_purchase_timestamp</th>\n",
       "      <th>payment_sequential</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>price</th>\n",
       "      <th>payment_value</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_photos_qty</th>\n",
       "      <th>product_name_lenght</th>\n",
       "      <th>volume</th>\n",
       "      <th>shipping_limit_date</th>\n",
       "      <th>order_delivered_carrier_date</th>\n",
       "      <th>shipping_time</th>\n",
       "      <th>review_delay</th>\n",
       "      <th>review_answer_delay</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0000366f3b9a7992bf8c76cfdf3221e2</th>\n",
       "      <td>2018-05-17</td>\n",
       "      <td>2018-05-21 13:44:48</td>\n",
       "      <td>10.533819</td>\n",
       "      <td>6.400914</td>\n",
       "      <td>0.010313</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-05-10 10:56:27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.484907</td>\n",
       "      <td>4.866765</td>\n",
       "      <td>141.90</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7616.0</td>\n",
       "      <td>2018-05-15 11:11:18</td>\n",
       "      <td>2018-05-12 08:18:00</td>\n",
       "      <td>3.120347</td>\n",
       "      <td>6.544132</td>\n",
       "      <td>4.572778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000b849f77a49e4a4ce2b2a4ca5be3f</th>\n",
       "      <td>2018-05-11</td>\n",
       "      <td>2018-05-11 22:29:25</td>\n",
       "      <td>7.232130</td>\n",
       "      <td>2.984005</td>\n",
       "      <td>0.301586</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-05-07 11:11:27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.115050</td>\n",
       "      <td>2.939162</td>\n",
       "      <td>27.19</td>\n",
       "      <td>1.0</td>\n",
       "      <td>375.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>5148.0</td>\n",
       "      <td>2018-05-11 17:56:33</td>\n",
       "      <td>2018-05-09 12:18:00</td>\n",
       "      <td>2.235104</td>\n",
       "      <td>3.533715</td>\n",
       "      <td>0.937095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f46a3911fa3c0805444483337064</th>\n",
       "      <td>2017-04-06</td>\n",
       "      <td>2017-04-07 01:19:44</td>\n",
       "      <td>27.121493</td>\n",
       "      <td>25.731759</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2017-03-10 21:05:03</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.846071</td>\n",
       "      <td>4.234107</td>\n",
       "      <td>86.22</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>49.0</td>\n",
       "      <td>43750.0</td>\n",
       "      <td>2017-03-15 21:05:03</td>\n",
       "      <td>2017-03-13 12:58:30</td>\n",
       "      <td>2.337882</td>\n",
       "      <td>26.121493</td>\n",
       "      <td>1.055370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f6ccb0745a6a4b88665a16c9f078</th>\n",
       "      <td>2017-11-02</td>\n",
       "      <td>2017-11-03 00:17:51</td>\n",
       "      <td>31.132442</td>\n",
       "      <td>20.023472</td>\n",
       "      <td>0.013611</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2017-10-12 20:29:41</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.869602</td>\n",
       "      <td>3.257712</td>\n",
       "      <td>43.62</td>\n",
       "      <td>1.0</td>\n",
       "      <td>150.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>1045.0</td>\n",
       "      <td>2017-10-18 21:49:17</td>\n",
       "      <td>2017-10-13 20:08:19</td>\n",
       "      <td>5.070116</td>\n",
       "      <td>20.146053</td>\n",
       "      <td>1.012396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0004aac84e0df4da2b147fca70cf8255</th>\n",
       "      <td>2017-11-28</td>\n",
       "      <td>2017-12-02 22:29:35</td>\n",
       "      <td>20.161898</td>\n",
       "      <td>13.126435</td>\n",
       "      <td>0.014699</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2017-11-14 19:45:42</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.826722</td>\n",
       "      <td>5.192957</td>\n",
       "      <td>196.89</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6050.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>528.0</td>\n",
       "      <td>2017-11-22 20:06:52</td>\n",
       "      <td>2017-11-16 19:52:10</td>\n",
       "      <td>6.010208</td>\n",
       "      <td>13.176597</td>\n",
       "      <td>4.937211</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 review_creation_date review_answer_timestamp  \\\n",
       "customer_unique_id                                                              \n",
       "0000366f3b9a7992bf8c76cfdf3221e2           2018-05-17     2018-05-21 13:44:48   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f           2018-05-11     2018-05-11 22:29:25   \n",
       "0000f46a3911fa3c0805444483337064           2017-04-06     2017-04-07 01:19:44   \n",
       "0000f6ccb0745a6a4b88665a16c9f078           2017-11-02     2017-11-03 00:17:51   \n",
       "0004aac84e0df4da2b147fca70cf8255           2017-11-28     2017-12-02 22:29:35   \n",
       "\n",
       "                                  estimated_delivery_time  delivery_time  \\\n",
       "customer_unique_id                                                         \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                10.533819       6.400914   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 7.232130       2.984005   \n",
       "0000f46a3911fa3c0805444483337064                27.121493      25.731759   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                31.132442      20.023472   \n",
       "0004aac84e0df4da2b147fca70cf8255                20.161898      13.126435   \n",
       "\n",
       "                                  approval_time  purchase_year  \\\n",
       "customer_unique_id                                               \n",
       "0000366f3b9a7992bf8c76cfdf3221e2       0.010313         2018.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f       0.301586         2018.0   \n",
       "0000f46a3911fa3c0805444483337064       0.000000         2017.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078       0.013611         2017.0   \n",
       "0004aac84e0df4da2b147fca70cf8255       0.014699         2017.0   \n",
       "\n",
       "                                  purchase_month order_purchase_timestamp  \\\n",
       "customer_unique_id                                                          \n",
       "0000366f3b9a7992bf8c76cfdf3221e2             5.0      2018-05-10 10:56:27   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f             5.0      2018-05-07 11:11:27   \n",
       "0000f46a3911fa3c0805444483337064             3.0      2017-03-10 21:05:03   \n",
       "0000f6ccb0745a6a4b88665a16c9f078            10.0      2017-10-12 20:29:41   \n",
       "0004aac84e0df4da2b147fca70cf8255            11.0      2017-11-14 19:45:42   \n",
       "\n",
       "                                  payment_sequential  freight_value     price  \\\n",
       "customer_unique_id                                                              \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                 1.0       2.484907  4.866765   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 1.0       2.115050  2.939162   \n",
       "0000f46a3911fa3c0805444483337064                 1.0       2.846071  4.234107   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                 1.0       2.869602  3.257712   \n",
       "0004aac84e0df4da2b147fca70cf8255                 1.0       2.826722  5.192957   \n",
       "\n",
       "                                  payment_value  order_item_id  \\\n",
       "customer_unique_id                                               \n",
       "0000366f3b9a7992bf8c76cfdf3221e2         141.90            1.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f          27.19            1.0   \n",
       "0000f46a3911fa3c0805444483337064          86.22            1.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078          43.62            1.0   \n",
       "0004aac84e0df4da2b147fca70cf8255         196.89            1.0   \n",
       "\n",
       "                                  product_weight_g  product_photos_qty  \\\n",
       "customer_unique_id                                                       \n",
       "0000366f3b9a7992bf8c76cfdf3221e2            1500.0                 1.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f             375.0                 1.0   \n",
       "0000f46a3911fa3c0805444483337064            1500.0                 3.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078             150.0                 5.0   \n",
       "0004aac84e0df4da2b147fca70cf8255            6050.0                 3.0   \n",
       "\n",
       "                                  product_name_lenght   volume  \\\n",
       "customer_unique_id                                               \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                 60.0   7616.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 56.0   5148.0   \n",
       "0000f46a3911fa3c0805444483337064                 49.0  43750.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                 43.0   1045.0   \n",
       "0004aac84e0df4da2b147fca70cf8255                 58.0    528.0   \n",
       "\n",
       "                                 shipping_limit_date  \\\n",
       "customer_unique_id                                     \n",
       "0000366f3b9a7992bf8c76cfdf3221e2 2018-05-15 11:11:18   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f 2018-05-11 17:56:33   \n",
       "0000f46a3911fa3c0805444483337064 2017-03-15 21:05:03   \n",
       "0000f6ccb0745a6a4b88665a16c9f078 2017-10-18 21:49:17   \n",
       "0004aac84e0df4da2b147fca70cf8255 2017-11-22 20:06:52   \n",
       "\n",
       "                                 order_delivered_carrier_date  shipping_time  \\\n",
       "customer_unique_id                                                             \n",
       "0000366f3b9a7992bf8c76cfdf3221e2          2018-05-12 08:18:00       3.120347   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f          2018-05-09 12:18:00       2.235104   \n",
       "0000f46a3911fa3c0805444483337064          2017-03-13 12:58:30       2.337882   \n",
       "0000f6ccb0745a6a4b88665a16c9f078          2017-10-13 20:08:19       5.070116   \n",
       "0004aac84e0df4da2b147fca70cf8255          2017-11-16 19:52:10       6.010208   \n",
       "\n",
       "                                  review_delay  review_answer_delay  \n",
       "customer_unique_id                                                   \n",
       "0000366f3b9a7992bf8c76cfdf3221e2      6.544132             4.572778  \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f      3.533715             0.937095  \n",
       "0000f46a3911fa3c0805444483337064     26.121493             1.055370  \n",
       "0000f6ccb0745a6a4b88665a16c9f078     20.146053             1.012396  \n",
       "0004aac84e0df4da2b147fca70cf8255     13.176597             4.937211  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "753e361c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 96096 entries, 0000366f3b9a7992bf8c76cfdf3221e2 to ffffd2657e2aad2907e67c3e9daecbeb\n",
      "Data columns (total 22 columns):\n",
      " #   Column                        Non-Null Count  Dtype         \n",
      "---  ------                        --------------  -----         \n",
      " 0   review_creation_date          95380 non-null  datetime64[ns]\n",
      " 1   review_answer_timestamp       95380 non-null  datetime64[ns]\n",
      " 2   estimated_delivery_time       95997 non-null  float64       \n",
      " 3   delivery_time                 93343 non-null  float64       \n",
      " 4   approval_time                 95997 non-null  float64       \n",
      " 5   purchase_year                 96096 non-null  float64       \n",
      " 6   purchase_month                96096 non-null  float64       \n",
      " 7   order_purchase_timestamp      96096 non-null  datetime64[ns]\n",
      " 8   payment_sequential            96096 non-null  float64       \n",
      " 9   freight_value                 96096 non-null  float64       \n",
      " 10  price                         96096 non-null  float64       \n",
      " 11  payment_value                 96096 non-null  float64       \n",
      " 12  order_item_id                 96096 non-null  float64       \n",
      " 13  product_weight_g              95407 non-null  float64       \n",
      " 14  product_photos_qty            94108 non-null  float64       \n",
      " 15  product_name_lenght           94108 non-null  float64       \n",
      " 16  volume                        95407 non-null  float64       \n",
      " 17  shipping_limit_date           95420 non-null  datetime64[ns]\n",
      " 18  order_delivered_carrier_date  94466 non-null  datetime64[ns]\n",
      " 19  shipping_time                 94465 non-null  float64       \n",
      " 20  review_delay                  95380 non-null  float64       \n",
      " 21  review_answer_delay           95380 non-null  float64       \n",
      "dtypes: datetime64[ns](5), float64(17)\n",
      "memory usage: 16.9+ MB\n"
     ]
    }
   ],
   "source": [
    "model.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "0a66830a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "price                           0.000000\n",
       "purchase_year                   0.000000\n",
       "purchase_month                  0.000000\n",
       "order_purchase_timestamp        0.000000\n",
       "payment_sequential              0.000000\n",
       "freight_value                   0.000000\n",
       "payment_value                   0.000000\n",
       "order_item_id                   0.000000\n",
       "estimated_delivery_time         0.083093\n",
       "approval_time                   0.083093\n",
       "shipping_limit_date             0.567385\n",
       "volume                          0.578297\n",
       "product_weight_g                0.578297\n",
       "review_creation_date            0.600959\n",
       "review_answer_timestamp         0.600959\n",
       "review_delay                    0.600959\n",
       "review_answer_delay             0.600959\n",
       "order_delivered_carrier_date    1.368104\n",
       "shipping_time                   1.368943\n",
       "product_photos_qty              1.668583\n",
       "product_name_lenght             1.668583\n",
       "delivery_time                   2.310669\n",
       "dtype: float64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentage of nan for each column of data\n",
    "df_missing = model.isna().sum().sort_values()/len(data)*100\n",
    "df_missing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "02b21f6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "sample = model.join(rfm[['RecencyScore','FrequencyScore','MonetaryScore']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "497c265a",
   "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>review_creation_date</th>\n",
       "      <th>review_answer_timestamp</th>\n",
       "      <th>estimated_delivery_time</th>\n",
       "      <th>delivery_time</th>\n",
       "      <th>approval_time</th>\n",
       "      <th>purchase_year</th>\n",
       "      <th>purchase_month</th>\n",
       "      <th>order_purchase_timestamp</th>\n",
       "      <th>payment_sequential</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>price</th>\n",
       "      <th>payment_value</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_photos_qty</th>\n",
       "      <th>product_name_lenght</th>\n",
       "      <th>volume</th>\n",
       "      <th>shipping_limit_date</th>\n",
       "      <th>order_delivered_carrier_date</th>\n",
       "      <th>shipping_time</th>\n",
       "      <th>review_delay</th>\n",
       "      <th>review_answer_delay</th>\n",
       "      <th>RecencyScore</th>\n",
       "      <th>FrequencyScore</th>\n",
       "      <th>MonetaryScore</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0000366f3b9a7992bf8c76cfdf3221e2</th>\n",
       "      <td>2018-05-17</td>\n",
       "      <td>2018-05-21 13:44:48</td>\n",
       "      <td>10.533819</td>\n",
       "      <td>6.400914</td>\n",
       "      <td>0.010313</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-05-10 10:56:27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.484907</td>\n",
       "      <td>4.866765</td>\n",
       "      <td>141.90</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7616.0</td>\n",
       "      <td>2018-05-15 11:11:18</td>\n",
       "      <td>2018-05-12 08:18:00</td>\n",
       "      <td>3.120347</td>\n",
       "      <td>6.544132</td>\n",
       "      <td>4.572778</td>\n",
       "      <td>160</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.955123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000b849f77a49e4a4ce2b2a4ca5be3f</th>\n",
       "      <td>2018-05-11</td>\n",
       "      <td>2018-05-11 22:29:25</td>\n",
       "      <td>7.232130</td>\n",
       "      <td>2.984005</td>\n",
       "      <td>0.301586</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-05-07 11:11:27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.115050</td>\n",
       "      <td>2.939162</td>\n",
       "      <td>27.19</td>\n",
       "      <td>1.0</td>\n",
       "      <td>375.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>5148.0</td>\n",
       "      <td>2018-05-11 17:56:33</td>\n",
       "      <td>2018-05-09 12:18:00</td>\n",
       "      <td>2.235104</td>\n",
       "      <td>3.533715</td>\n",
       "      <td>0.937095</td>\n",
       "      <td>163</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.302849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f46a3911fa3c0805444483337064</th>\n",
       "      <td>2017-04-06</td>\n",
       "      <td>2017-04-07 01:19:44</td>\n",
       "      <td>27.121493</td>\n",
       "      <td>25.731759</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2017-03-10 21:05:03</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.846071</td>\n",
       "      <td>4.234107</td>\n",
       "      <td>86.22</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>49.0</td>\n",
       "      <td>43750.0</td>\n",
       "      <td>2017-03-15 21:05:03</td>\n",
       "      <td>2017-03-13 12:58:30</td>\n",
       "      <td>2.337882</td>\n",
       "      <td>26.121493</td>\n",
       "      <td>1.055370</td>\n",
       "      <td>585</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.456902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f6ccb0745a6a4b88665a16c9f078</th>\n",
       "      <td>2017-11-02</td>\n",
       "      <td>2017-11-03 00:17:51</td>\n",
       "      <td>31.132442</td>\n",
       "      <td>20.023472</td>\n",
       "      <td>0.013611</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2017-10-12 20:29:41</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.869602</td>\n",
       "      <td>3.257712</td>\n",
       "      <td>43.62</td>\n",
       "      <td>1.0</td>\n",
       "      <td>150.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>1045.0</td>\n",
       "      <td>2017-10-18 21:49:17</td>\n",
       "      <td>2017-10-13 20:08:19</td>\n",
       "      <td>5.070116</td>\n",
       "      <td>20.146053</td>\n",
       "      <td>1.012396</td>\n",
       "      <td>369</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.775516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0004aac84e0df4da2b147fca70cf8255</th>\n",
       "      <td>2017-11-28</td>\n",
       "      <td>2017-12-02 22:29:35</td>\n",
       "      <td>20.161898</td>\n",
       "      <td>13.126435</td>\n",
       "      <td>0.014699</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2017-11-14 19:45:42</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.826722</td>\n",
       "      <td>5.192957</td>\n",
       "      <td>196.89</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6050.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>528.0</td>\n",
       "      <td>2017-11-22 20:06:52</td>\n",
       "      <td>2017-11-16 19:52:10</td>\n",
       "      <td>6.010208</td>\n",
       "      <td>13.176597</td>\n",
       "      <td>4.937211</td>\n",
       "      <td>336</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.282645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fffcf5a5ff07b0908bd4e2dbc735a684</th>\n",
       "      <td>2017-07-07</td>\n",
       "      <td>2017-07-11 11:20:05</td>\n",
       "      <td>54.114398</td>\n",
       "      <td>27.504676</td>\n",
       "      <td>0.010185</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2017-06-08 21:00:36</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.031093</td>\n",
       "      <td>13.313314</td>\n",
       "      <td>4134.84</td>\n",
       "      <td>3.0</td>\n",
       "      <td>30000.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>251625.0</td>\n",
       "      <td>2017-06-25 21:15:16</td>\n",
       "      <td>2017-06-23 13:03:27</td>\n",
       "      <td>2.341539</td>\n",
       "      <td>28.124583</td>\n",
       "      <td>4.472280</td>\n",
       "      <td>495</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>8.327204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fffea47cd6d3cc0a88bd621562a9d061</th>\n",
       "      <td>2018-01-10</td>\n",
       "      <td>2018-01-15 11:51:59</td>\n",
       "      <td>32.155324</td>\n",
       "      <td>30.091667</td>\n",
       "      <td>0.005833</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2017-12-10 20:07:56</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.980111</td>\n",
       "      <td>4.172694</td>\n",
       "      <td>84.58</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1720.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>32076.0</td>\n",
       "      <td>2017-12-14 20:16:20</td>\n",
       "      <td>2017-12-21 16:27:01</td>\n",
       "      <td>-6.840752</td>\n",
       "      <td>30.161157</td>\n",
       "      <td>5.494433</td>\n",
       "      <td>310</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.437698</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffff371b4d645b6ecea244b27531430a</th>\n",
       "      <td>2017-02-23</td>\n",
       "      <td>2017-02-25 15:03:21</td>\n",
       "      <td>44.331435</td>\n",
       "      <td>14.862731</td>\n",
       "      <td>0.009352</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2017-02-07 15:49:16</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.116178</td>\n",
       "      <td>4.498698</td>\n",
       "      <td>112.46</td>\n",
       "      <td>1.0</td>\n",
       "      <td>600.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>3315.0</td>\n",
       "      <td>2017-02-11 15:49:16</td>\n",
       "      <td>2017-02-08 10:04:37</td>\n",
       "      <td>3.239340</td>\n",
       "      <td>15.340787</td>\n",
       "      <td>2.627326</td>\n",
       "      <td>617</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.722598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffff5962728ec6157033ef9805bacc48</th>\n",
       "      <td>2018-05-15</td>\n",
       "      <td>2018-05-17 15:16:20</td>\n",
       "      <td>25.334178</td>\n",
       "      <td>11.830313</td>\n",
       "      <td>0.028542</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2018-05-02 15:17:41</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.927989</td>\n",
       "      <td>4.744932</td>\n",
       "      <td>133.69</td>\n",
       "      <td>1.0</td>\n",
       "      <td>444.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>4480.0</td>\n",
       "      <td>2018-05-08 15:58:47</td>\n",
       "      <td>2018-05-03 16:45:00</td>\n",
       "      <td>4.967905</td>\n",
       "      <td>12.362720</td>\n",
       "      <td>2.636343</td>\n",
       "      <td>168</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.895524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffffd2657e2aad2907e67c3e9daecbeb</th>\n",
       "      <td>2017-05-11</td>\n",
       "      <td>2017-05-13 10:55:24</td>\n",
       "      <td>22.364433</td>\n",
       "      <td>6.762905</td>\n",
       "      <td>0.789213</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2017-05-02 20:18:45</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.678965</td>\n",
       "      <td>4.042876</td>\n",
       "      <td>71.56</td>\n",
       "      <td>1.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>2178.0</td>\n",
       "      <td>2017-05-10 15:15:13</td>\n",
       "      <td>2017-05-04 12:05:22</td>\n",
       "      <td>6.131840</td>\n",
       "      <td>8.153646</td>\n",
       "      <td>2.455139</td>\n",
       "      <td>532</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.270536</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>96096 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 review_creation_date review_answer_timestamp  \\\n",
       "customer_unique_id                                                              \n",
       "0000366f3b9a7992bf8c76cfdf3221e2           2018-05-17     2018-05-21 13:44:48   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f           2018-05-11     2018-05-11 22:29:25   \n",
       "0000f46a3911fa3c0805444483337064           2017-04-06     2017-04-07 01:19:44   \n",
       "0000f6ccb0745a6a4b88665a16c9f078           2017-11-02     2017-11-03 00:17:51   \n",
       "0004aac84e0df4da2b147fca70cf8255           2017-11-28     2017-12-02 22:29:35   \n",
       "...                                               ...                     ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684           2017-07-07     2017-07-11 11:20:05   \n",
       "fffea47cd6d3cc0a88bd621562a9d061           2018-01-10     2018-01-15 11:51:59   \n",
       "ffff371b4d645b6ecea244b27531430a           2017-02-23     2017-02-25 15:03:21   \n",
       "ffff5962728ec6157033ef9805bacc48           2018-05-15     2018-05-17 15:16:20   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb           2017-05-11     2017-05-13 10:55:24   \n",
       "\n",
       "                                  estimated_delivery_time  delivery_time  \\\n",
       "customer_unique_id                                                         \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                10.533819       6.400914   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 7.232130       2.984005   \n",
       "0000f46a3911fa3c0805444483337064                27.121493      25.731759   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                31.132442      20.023472   \n",
       "0004aac84e0df4da2b147fca70cf8255                20.161898      13.126435   \n",
       "...                                                   ...            ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684                54.114398      27.504676   \n",
       "fffea47cd6d3cc0a88bd621562a9d061                32.155324      30.091667   \n",
       "ffff371b4d645b6ecea244b27531430a                44.331435      14.862731   \n",
       "ffff5962728ec6157033ef9805bacc48                25.334178      11.830313   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb                22.364433       6.762905   \n",
       "\n",
       "                                  approval_time  purchase_year  \\\n",
       "customer_unique_id                                               \n",
       "0000366f3b9a7992bf8c76cfdf3221e2       0.010313         2018.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f       0.301586         2018.0   \n",
       "0000f46a3911fa3c0805444483337064       0.000000         2017.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078       0.013611         2017.0   \n",
       "0004aac84e0df4da2b147fca70cf8255       0.014699         2017.0   \n",
       "...                                         ...            ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684       0.010185         2017.0   \n",
       "fffea47cd6d3cc0a88bd621562a9d061       0.005833         2017.0   \n",
       "ffff371b4d645b6ecea244b27531430a       0.009352         2017.0   \n",
       "ffff5962728ec6157033ef9805bacc48       0.028542         2018.0   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb       0.789213         2017.0   \n",
       "\n",
       "                                  purchase_month order_purchase_timestamp  \\\n",
       "customer_unique_id                                                          \n",
       "0000366f3b9a7992bf8c76cfdf3221e2             5.0      2018-05-10 10:56:27   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f             5.0      2018-05-07 11:11:27   \n",
       "0000f46a3911fa3c0805444483337064             3.0      2017-03-10 21:05:03   \n",
       "0000f6ccb0745a6a4b88665a16c9f078            10.0      2017-10-12 20:29:41   \n",
       "0004aac84e0df4da2b147fca70cf8255            11.0      2017-11-14 19:45:42   \n",
       "...                                          ...                      ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684             6.0      2017-06-08 21:00:36   \n",
       "fffea47cd6d3cc0a88bd621562a9d061            12.0      2017-12-10 20:07:56   \n",
       "ffff371b4d645b6ecea244b27531430a             2.0      2017-02-07 15:49:16   \n",
       "ffff5962728ec6157033ef9805bacc48             5.0      2018-05-02 15:17:41   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb             5.0      2017-05-02 20:18:45   \n",
       "\n",
       "                                  payment_sequential  freight_value  \\\n",
       "customer_unique_id                                                    \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                 1.0       2.484907   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 1.0       2.115050   \n",
       "0000f46a3911fa3c0805444483337064                 1.0       2.846071   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                 1.0       2.869602   \n",
       "0004aac84e0df4da2b147fca70cf8255                 1.0       2.826722   \n",
       "...                                              ...            ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684                 2.0      11.031093   \n",
       "fffea47cd6d3cc0a88bd621562a9d061                 1.0       2.980111   \n",
       "ffff371b4d645b6ecea244b27531430a                 1.0       3.116178   \n",
       "ffff5962728ec6157033ef9805bacc48                 1.0       2.927989   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb                 1.0       2.678965   \n",
       "\n",
       "                                      price  payment_value  order_item_id  \\\n",
       "customer_unique_id                                                          \n",
       "0000366f3b9a7992bf8c76cfdf3221e2   4.866765         141.90            1.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f   2.939162          27.19            1.0   \n",
       "0000f46a3911fa3c0805444483337064   4.234107          86.22            1.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078   3.257712          43.62            1.0   \n",
       "0004aac84e0df4da2b147fca70cf8255   5.192957         196.89            1.0   \n",
       "...                                     ...            ...            ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684  13.313314        4134.84            3.0   \n",
       "fffea47cd6d3cc0a88bd621562a9d061   4.172694          84.58            1.0   \n",
       "ffff371b4d645b6ecea244b27531430a   4.498698         112.46            1.0   \n",
       "ffff5962728ec6157033ef9805bacc48   4.744932         133.69            1.0   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb   4.042876          71.56            1.0   \n",
       "\n",
       "                                  product_weight_g  product_photos_qty  \\\n",
       "customer_unique_id                                                       \n",
       "0000366f3b9a7992bf8c76cfdf3221e2            1500.0                 1.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f             375.0                 1.0   \n",
       "0000f46a3911fa3c0805444483337064            1500.0                 3.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078             150.0                 5.0   \n",
       "0004aac84e0df4da2b147fca70cf8255            6050.0                 3.0   \n",
       "...                                            ...                 ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684           30000.0                 1.0   \n",
       "fffea47cd6d3cc0a88bd621562a9d061            1720.0                 2.0   \n",
       "ffff371b4d645b6ecea244b27531430a             600.0                 5.0   \n",
       "ffff5962728ec6157033ef9805bacc48             444.0                 3.0   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb             400.0                 2.0   \n",
       "\n",
       "                                  product_name_lenght    volume  \\\n",
       "customer_unique_id                                                \n",
       "0000366f3b9a7992bf8c76cfdf3221e2                 60.0    7616.0   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f                 56.0    5148.0   \n",
       "0000f46a3911fa3c0805444483337064                 49.0   43750.0   \n",
       "0000f6ccb0745a6a4b88665a16c9f078                 43.0    1045.0   \n",
       "0004aac84e0df4da2b147fca70cf8255                 58.0     528.0   \n",
       "...                                               ...       ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684                 54.0  251625.0   \n",
       "fffea47cd6d3cc0a88bd621562a9d061                 37.0   32076.0   \n",
       "ffff371b4d645b6ecea244b27531430a                 57.0    3315.0   \n",
       "ffff5962728ec6157033ef9805bacc48                 35.0    4480.0   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb                 45.0    2178.0   \n",
       "\n",
       "                                 shipping_limit_date  \\\n",
       "customer_unique_id                                     \n",
       "0000366f3b9a7992bf8c76cfdf3221e2 2018-05-15 11:11:18   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f 2018-05-11 17:56:33   \n",
       "0000f46a3911fa3c0805444483337064 2017-03-15 21:05:03   \n",
       "0000f6ccb0745a6a4b88665a16c9f078 2017-10-18 21:49:17   \n",
       "0004aac84e0df4da2b147fca70cf8255 2017-11-22 20:06:52   \n",
       "...                                              ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684 2017-06-25 21:15:16   \n",
       "fffea47cd6d3cc0a88bd621562a9d061 2017-12-14 20:16:20   \n",
       "ffff371b4d645b6ecea244b27531430a 2017-02-11 15:49:16   \n",
       "ffff5962728ec6157033ef9805bacc48 2018-05-08 15:58:47   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb 2017-05-10 15:15:13   \n",
       "\n",
       "                                 order_delivered_carrier_date  shipping_time  \\\n",
       "customer_unique_id                                                             \n",
       "0000366f3b9a7992bf8c76cfdf3221e2          2018-05-12 08:18:00       3.120347   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f          2018-05-09 12:18:00       2.235104   \n",
       "0000f46a3911fa3c0805444483337064          2017-03-13 12:58:30       2.337882   \n",
       "0000f6ccb0745a6a4b88665a16c9f078          2017-10-13 20:08:19       5.070116   \n",
       "0004aac84e0df4da2b147fca70cf8255          2017-11-16 19:52:10       6.010208   \n",
       "...                                                       ...            ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684          2017-06-23 13:03:27       2.341539   \n",
       "fffea47cd6d3cc0a88bd621562a9d061          2017-12-21 16:27:01      -6.840752   \n",
       "ffff371b4d645b6ecea244b27531430a          2017-02-08 10:04:37       3.239340   \n",
       "ffff5962728ec6157033ef9805bacc48          2018-05-03 16:45:00       4.967905   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb          2017-05-04 12:05:22       6.131840   \n",
       "\n",
       "                                  review_delay  review_answer_delay  \\\n",
       "customer_unique_id                                                    \n",
       "0000366f3b9a7992bf8c76cfdf3221e2      6.544132             4.572778   \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f      3.533715             0.937095   \n",
       "0000f46a3911fa3c0805444483337064     26.121493             1.055370   \n",
       "0000f6ccb0745a6a4b88665a16c9f078     20.146053             1.012396   \n",
       "0004aac84e0df4da2b147fca70cf8255     13.176597             4.937211   \n",
       "...                                        ...                  ...   \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684     28.124583             4.472280   \n",
       "fffea47cd6d3cc0a88bd621562a9d061     30.161157             5.494433   \n",
       "ffff371b4d645b6ecea244b27531430a     15.340787             2.627326   \n",
       "ffff5962728ec6157033ef9805bacc48     12.362720             2.636343   \n",
       "ffffd2657e2aad2907e67c3e9daecbeb      8.153646             2.455139   \n",
       "\n",
       "                                  RecencyScore  FrequencyScore  MonetaryScore  \n",
       "customer_unique_id                                                             \n",
       "0000366f3b9a7992bf8c76cfdf3221e2           160        0.000000       4.955123  \n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f           163        0.000000       3.302849  \n",
       "0000f46a3911fa3c0805444483337064           585        0.000000       4.456902  \n",
       "0000f6ccb0745a6a4b88665a16c9f078           369        0.000000       3.775516  \n",
       "0004aac84e0df4da2b147fca70cf8255           336        0.000000       5.282645  \n",
       "...                                        ...             ...            ...  \n",
       "fffcf5a5ff07b0908bd4e2dbc735a684           495        0.693147       8.327204  \n",
       "fffea47cd6d3cc0a88bd621562a9d061           310        0.000000       4.437698  \n",
       "ffff371b4d645b6ecea244b27531430a           617        0.000000       4.722598  \n",
       "ffff5962728ec6157033ef9805bacc48           168        0.000000       4.895524  \n",
       "ffffd2657e2aad2907e67c3e9daecbeb           532        0.000000       4.270536  \n",
       "\n",
       "[96096 rows x 25 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db471463",
   "metadata": {},
   "source": [
    "## Echantillonnage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "2e92761a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 5% selection of a sample \n",
    "sample = rfm.sample(frac=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "6c2199b3",
   "metadata": {
    "scrolled": true
   },
   "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>RecencyScore</th>\n",
       "      <th>FrequencyScore</th>\n",
       "      <th>MonetaryScore</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>d0af5dec42f3bf160195f5bcb704fd01</th>\n",
       "      <td>412</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>3.672242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>abf60c6ebc739e6f4e4319a9044073af</th>\n",
       "      <td>323</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.496348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4d50764ac486be215e204a03f2ab11c8</th>\n",
       "      <td>271</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.057073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67f5420469ecf30af33c56afc4c82c68</th>\n",
       "      <td>288</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.637849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e5d380f1cd5c220fdab47e0245471013</th>\n",
       "      <td>501</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.667018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>eca61ae704cc5ecddd7a6e558b6f04b1</th>\n",
       "      <td>200</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>8.099585</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49c6b1ee895e59043bcd5e2135f08355</th>\n",
       "      <td>145</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>5.394899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>adddad3f8f29debdcafc57d768c540ac</th>\n",
       "      <td>490</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.496202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7c50190d7200a712b3efbf74dcda62a9</th>\n",
       "      <td>132</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.123580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ae86daa72f710105fb92e3843e356425</th>\n",
       "      <td>103</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.471867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8f3111ce59c5dc5de1f424c61c2bd9a8</th>\n",
       "      <td>83</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.114094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>09b496dcae2c6c126812d20369b44bf8</th>\n",
       "      <td>326</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.850148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57d91ead18f17efa88295e6a1a8e88f1</th>\n",
       "      <td>259</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.391398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2f3ffd9f28e0a140401a51e8ec5ec4d6</th>\n",
       "      <td>74</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>5.721491</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a48dc9ff5e16f55336756967c43d7431</th>\n",
       "      <td>589</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.449685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9a27c84d671dfcce14be5e6a9af0f3f8</th>\n",
       "      <td>462</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.313748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dbcefdf152f6054ebca8ca78edabd755</th>\n",
       "      <td>467</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.770176</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7ed95317b0006c3104d204efb64848de</th>\n",
       "      <td>354</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.971235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2279cf489dcf3b1d919069199fd70387</th>\n",
       "      <td>76</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.888995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b6488c3c9ac8506f2bfcd968e5622891</th>\n",
       "      <td>157</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.800737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c297e9bcd0395171a527a01ecfe8745d</th>\n",
       "      <td>493</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.446643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121d81c074edf74f0338b8688c1e8f6f</th>\n",
       "      <td>236</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>5.369335</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5736be301e11f2ce0b0aa6fcc03b52c8</th>\n",
       "      <td>186</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>5.272384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4c2648c5850ac8a3f7105333ef641e6b</th>\n",
       "      <td>88</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.813307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6f780c5a7fe8b1bd7d629c7a5851d5bc</th>\n",
       "      <td>130</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.699038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0f19406817348f175890e311daa0f12a</th>\n",
       "      <td>503</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.432805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b0765213c38298fb584fee79eb17099e</th>\n",
       "      <td>403</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.992946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1e3c7109171cafe8b09225708b11c383</th>\n",
       "      <td>384</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.174387</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78459b88a80c9a5b01301a41c304f21c</th>\n",
       "      <td>261</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.976803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55f8fb3a399773c4a452881895330a84</th>\n",
       "      <td>176</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.307029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>845242d03a9bf44fb38894d7465c407e</th>\n",
       "      <td>264</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.149622</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2989d990cbaf75e60b262583cacda3d4</th>\n",
       "      <td>531</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.664587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ae81f7525461d59870e615dcecd126b7</th>\n",
       "      <td>183</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.663911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c25c0de4341701c95fb2675d9d82d414</th>\n",
       "      <td>246</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.739876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7dea8532b362518e0bf338eafdab0131</th>\n",
       "      <td>72</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.286616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>06f59c70f97379ce2f67075dd6ae26df</th>\n",
       "      <td>166</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.362079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>319e86b5a7cbc42a85be7e1e452f337a</th>\n",
       "      <td>396</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.731462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52387f8339f0404a18ce6c12139bcd6e</th>\n",
       "      <td>141</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.668802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e4ef5faceeaffad29e3878e8a930953b</th>\n",
       "      <td>550</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.966196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8a0c0a5a27f03eb1ee71c9ed1985a329</th>\n",
       "      <td>260</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.987507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8d434defc893fd9ac17fabd1462cbce3</th>\n",
       "      <td>154</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.512612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bb58670190dba4e9b320f84cb98317a3</th>\n",
       "      <td>489</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>5.349723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a7c14e93bbe5b0fa949c03c3c1eecb26</th>\n",
       "      <td>325</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.212128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>944b31e72098bf995392fea4debf05a6</th>\n",
       "      <td>301</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>3.529297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dee19bd0f0519deb0b5cb50e2f1da9dc</th>\n",
       "      <td>251</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.094118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>006bbb235068a01d2e3e47fa09ccf53c</th>\n",
       "      <td>243</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.526361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b6ea04eb4b04276bb4e9731099c85443</th>\n",
       "      <td>258</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.011687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36cff82c2ab27c1e3bfee5d0b4fc4e6f</th>\n",
       "      <td>288</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.596634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8fb08dc39120bb8e449977b2a1ff880c</th>\n",
       "      <td>286</td>\n",
       "      <td>0.693147</td>\n",
       "      <td>5.384679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81814a8a85ef3b0cace32bf039a7a31c</th>\n",
       "      <td>70</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.587617</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  RecencyScore  FrequencyScore  MonetaryScore\n",
       "customer_unique_id                                                           \n",
       "d0af5dec42f3bf160195f5bcb704fd01           412        0.693147       3.672242\n",
       "abf60c6ebc739e6f4e4319a9044073af           323        0.000000       5.496348\n",
       "4d50764ac486be215e204a03f2ab11c8           271        0.000000       5.057073\n",
       "67f5420469ecf30af33c56afc4c82c68           288        0.000000       3.637849\n",
       "e5d380f1cd5c220fdab47e0245471013           501        0.000000       4.667018\n",
       "eca61ae704cc5ecddd7a6e558b6f04b1           200        1.098612       8.099585\n",
       "49c6b1ee895e59043bcd5e2135f08355           145        0.693147       5.394899\n",
       "adddad3f8f29debdcafc57d768c540ac           490        0.000000       6.496202\n",
       "7c50190d7200a712b3efbf74dcda62a9           132        0.000000       4.123580\n",
       "ae86daa72f710105fb92e3843e356425           103        0.000000       4.471867\n",
       "8f3111ce59c5dc5de1f424c61c2bd9a8            83        0.000000       5.114094\n",
       "09b496dcae2c6c126812d20369b44bf8           326        0.000000       3.850148\n",
       "57d91ead18f17efa88295e6a1a8e88f1           259        0.000000       5.391398\n",
       "2f3ffd9f28e0a140401a51e8ec5ec4d6            74        1.386294       5.721491\n",
       "a48dc9ff5e16f55336756967c43d7431           589        0.000000       4.449685\n",
       "9a27c84d671dfcce14be5e6a9af0f3f8           462        0.000000       4.313748\n",
       "dbcefdf152f6054ebca8ca78edabd755           467        0.000000       4.770176\n",
       "7ed95317b0006c3104d204efb64848de           354        0.000000       3.971235\n",
       "2279cf489dcf3b1d919069199fd70387            76        0.000000       4.888995\n",
       "b6488c3c9ac8506f2bfcd968e5622891           157        0.000000       4.800737\n",
       "c297e9bcd0395171a527a01ecfe8745d           493        0.000000       4.446643\n",
       "121d81c074edf74f0338b8688c1e8f6f           236        0.693147       5.369335\n",
       "5736be301e11f2ce0b0aa6fcc03b52c8           186        0.693147       5.272384\n",
       "4c2648c5850ac8a3f7105333ef641e6b            88        0.000000       3.813307\n",
       "6f780c5a7fe8b1bd7d629c7a5851d5bc           130        0.000000       5.699038\n",
       "0f19406817348f175890e311daa0f12a           503        0.000000       5.432805\n",
       "b0765213c38298fb584fee79eb17099e           403        0.000000       4.992946\n",
       "1e3c7109171cafe8b09225708b11c383           384        0.000000       4.174387\n",
       "78459b88a80c9a5b01301a41c304f21c           261        0.000000       4.976803\n",
       "55f8fb3a399773c4a452881895330a84           176        0.000000       5.307029\n",
       "845242d03a9bf44fb38894d7465c407e           264        0.000000       4.149622\n",
       "2989d990cbaf75e60b262583cacda3d4           531        0.000000       3.664587\n",
       "ae81f7525461d59870e615dcecd126b7           183        0.000000       4.663911\n",
       "c25c0de4341701c95fb2675d9d82d414           246        0.000000       4.739876\n",
       "7dea8532b362518e0bf338eafdab0131            72        0.000000       4.286616\n",
       "06f59c70f97379ce2f67075dd6ae26df           166        0.000000       4.362079\n",
       "319e86b5a7cbc42a85be7e1e452f337a           396        0.000000       5.731462\n",
       "52387f8339f0404a18ce6c12139bcd6e           141        0.000000       4.668802\n",
       "e4ef5faceeaffad29e3878e8a930953b           550        0.000000       4.966196\n",
       "8a0c0a5a27f03eb1ee71c9ed1985a329           260        0.000000       5.987507\n",
       "8d434defc893fd9ac17fabd1462cbce3           154        0.000000       7.512612\n",
       "bb58670190dba4e9b320f84cb98317a3           489        1.609438       5.349723\n",
       "a7c14e93bbe5b0fa949c03c3c1eecb26           325        0.000000       4.212128\n",
       "944b31e72098bf995392fea4debf05a6           301        0.693147       3.529297\n",
       "dee19bd0f0519deb0b5cb50e2f1da9dc           251        0.000000       5.094118\n",
       "006bbb235068a01d2e3e47fa09ccf53c           243        0.000000       3.526361\n",
       "b6ea04eb4b04276bb4e9731099c85443           258        0.000000       4.011687\n",
       "36cff82c2ab27c1e3bfee5d0b4fc4e6f           288        0.000000       4.596634\n",
       "8fb08dc39120bb8e449977b2a1ff880c           286        0.693147       5.384679\n",
       "81814a8a85ef3b0cace32bf039a7a31c            70        0.000000       4.587617"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample.head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "8a1aa7ee",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 96096 entries, d0af5dec42f3bf160195f5bcb704fd01 to ec21b84d6332db46075bfb851000c59f\n",
      "Data columns (total 3 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   RecencyScore    96096 non-null  int64  \n",
      " 1   FrequencyScore  96096 non-null  float64\n",
      " 2   MonetaryScore   96093 non-null  float64\n",
      "dtypes: float64(2), int64(1)\n",
      "memory usage: 2.9+ MB\n"
     ]
    }
   ],
   "source": [
    "sample.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c55c7ce8",
   "metadata": {},
   "source": [
    "## Imputers, transformers and pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "ab870192",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imputers\n",
    "from sklearn.experimental import enable_iterative_imputer\n",
    "from sklearn.impute import IterativeImputer\n",
    "from sklearn.impute import SimpleImputer\n",
    "\n",
    "\n",
    "# Transformers\n",
    "from sklearn.decomposition import PCA\n",
    "from category_encoders.target_encoder import TargetEncoder\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.preprocessing import OrdinalEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import PowerTransformer\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.preprocessing import QuantileTransformer\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "\n",
    "# Pipeline\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.compose import TransformedTargetRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "36fa4edf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Separation into numerical and categorical features\n",
    "numerical_features = sample.select_dtypes(include=['int64', 'float64']).columns\n",
    "categorical_features = sample.select_dtypes(include=['object', 'bool']).columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a25ab0e4",
   "metadata": {},
   "source": [
    "pca = PCA(n_components=2)\n",
    "X_pca = pca.fit_transform(sample)\n",
    "total_explained_variance = pca.explained_variance_ratio_.cumsum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "bf1efc22",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# Preprocessing for numerical data\n",
    "numerical_transformer = Pipeline(steps=[\n",
    "    ('imputer', SimpleImputer(strategy='median')),\n",
    "    ('std_scaler', StandardScaler()),\n",
    "    #('PCA', PCA(n_components=3)),\n",
    "    #('std_scaler', RobustScaler()),\n",
    "    #('std_scaler', MinMaxScaler()),\n",
    "    #('QuantileTransformer', QuantileTransformer()),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "457a4106",
   "metadata": {},
   "outputs": [],
   "source": [
    "preprocessing_pipeline = Pipeline(steps=[('numerical_transformer', numerical_transformer)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "54de006a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['d0af5dec42f3bf160195f5bcb704fd01', 'abf60c6ebc739e6f4e4319a9044073af',\n",
      "       '4d50764ac486be215e204a03f2ab11c8', '67f5420469ecf30af33c56afc4c82c68',\n",
      "       'e5d380f1cd5c220fdab47e0245471013', 'eca61ae704cc5ecddd7a6e558b6f04b1',\n",
      "       '49c6b1ee895e59043bcd5e2135f08355', 'adddad3f8f29debdcafc57d768c540ac',\n",
      "       '7c50190d7200a712b3efbf74dcda62a9', 'ae86daa72f710105fb92e3843e356425',\n",
      "       ...\n",
      "       '01ad914ceb44eddfbf992570e6a5ef49', '3773db65dcc8cdff919dc782fa179380',\n",
      "       '420be7d41ee4292cb55d2616584b28f9', 'd5dea31916765489154e0f82c3e66bf7',\n",
      "       '1ae6159fb2d716e145ebaaa0224f0b46', '91a2d741f9510c7afbdcaa8b10c0050f',\n",
      "       'ad2bbf8d4d2af0537f9f5ff6f3ad8192', '38f700a297fe5d479187e8df810e5f66',\n",
      "       '0711ee90de545cb9be5c470614ad976b', 'ec21b84d6332db46075bfb851000c59f'],\n",
      "      dtype='object', name='customer_unique_id', length=96096)\n"
     ]
    }
   ],
   "source": [
    "index = sample.index\n",
    "print(index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2c71f865",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['RecencyScore', 'FrequencyScore', 'MonetaryScore'], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "columns = sample.columns\n",
    "print(columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "c11d57ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "d0af5dec42f3bf160195f5bcb704fd01    412\n",
       "abf60c6ebc739e6f4e4319a9044073af    323\n",
       "4d50764ac486be215e204a03f2ab11c8    271\n",
       "67f5420469ecf30af33c56afc4c82c68    288\n",
       "e5d380f1cd5c220fdab47e0245471013    501\n",
       "                                   ... \n",
       "91a2d741f9510c7afbdcaa8b10c0050f    385\n",
       "ad2bbf8d4d2af0537f9f5ff6f3ad8192    128\n",
       "38f700a297fe5d479187e8df810e5f66    205\n",
       "0711ee90de545cb9be5c470614ad976b    313\n",
       "ec21b84d6332db46075bfb851000c59f    490\n",
       "Name: RecencyScore, Length: 96096, dtype: int64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample.iloc[:,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "66f3a39f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "d0af5dec42f3bf160195f5bcb704fd01    0.693147\n",
       "abf60c6ebc739e6f4e4319a9044073af    0.000000\n",
       "4d50764ac486be215e204a03f2ab11c8    0.000000\n",
       "67f5420469ecf30af33c56afc4c82c68    0.000000\n",
       "e5d380f1cd5c220fdab47e0245471013    0.000000\n",
       "                                      ...   \n",
       "91a2d741f9510c7afbdcaa8b10c0050f    0.000000\n",
       "ad2bbf8d4d2af0537f9f5ff6f3ad8192    0.000000\n",
       "38f700a297fe5d479187e8df810e5f66    0.693147\n",
       "0711ee90de545cb9be5c470614ad976b    0.000000\n",
       "ec21b84d6332db46075bfb851000c59f    0.000000\n",
       "Name: FrequencyScore, Length: 96096, dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample.iloc[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "4452b18a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "d0af5dec42f3bf160195f5bcb704fd01    3.672242\n",
       "abf60c6ebc739e6f4e4319a9044073af    5.496348\n",
       "4d50764ac486be215e204a03f2ab11c8    5.057073\n",
       "67f5420469ecf30af33c56afc4c82c68    3.637849\n",
       "e5d380f1cd5c220fdab47e0245471013    4.667018\n",
       "                                      ...   \n",
       "91a2d741f9510c7afbdcaa8b10c0050f    4.144721\n",
       "ad2bbf8d4d2af0537f9f5ff6f3ad8192    3.253857\n",
       "38f700a297fe5d479187e8df810e5f66    5.586949\n",
       "0711ee90de545cb9be5c470614ad976b    4.798432\n",
       "ec21b84d6332db46075bfb851000c59f    4.219802\n",
       "Name: MonetaryScore, Length: 96096, dtype: float64"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample.iloc[:,2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "9a9a49d4",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "sample_scaled = preprocessing_pipeline.fit_transform(sample)\n",
    "#sample_scaled = pd.DataFrame(sample_scaled, index = sample.index)\n",
    "#sample_scaled = pd.DataFrame(sample_scaled,  index = sample.index, columns = sample.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "69574931",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1cd85ad13d0>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(5,5))\n",
    "plt.scatter(sample_scaled[:,0], sample_scaled[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "1e3fcda4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(96096, 3)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_scaled.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "5816ef9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#print(sample_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "718a6960",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(96096,)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_scaled[:,0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "43ab2101",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(96096,)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_scaled[:,1].shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a988db66",
   "metadata": {},
   "source": [
    "##  Métriques de Kmeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c02f413",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "from sklearn.metrics import davies_bouldin_score\n",
    "\n",
    "sse, slc, db = {}, {}, {}\n",
    "for i in range(2, 7):\n",
    "    kmeans = KMeans(init='k-means++', n_clusters=i, random_state=10, max_iter=400, n_init=50).fit(sample_scaled)\n",
    "    clusters = kmeans.labels_\n",
    "    sse[i] = kmeans.inertia_\n",
    "    slc[i] = silhouette_score(sample_scaled,clusters)\n",
    "    db[i] = davies_bouldin_score(sample_scaled,clusters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1036768f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Graphs for each metric\n",
    "\n",
    "fig, (ax0, ax1, ax2) = plt.subplots(1,3, figsize = (25, 5))\n",
    "\n",
    "ax0.plot(list(sse.keys()), list(sse.values()), color=\"blue\", marker='o')\n",
    "ax0.set_xlabel(\"number of cluster\")\n",
    "ax0.set_ylabel(\"inertia\")\n",
    "\n",
    "ax1.plot(list(slc.keys()), list(slc.values()), color=\"gold\", marker='o')\n",
    "ax1.set_xlabel(\"number of cluster\")\n",
    "ax1.set_ylabel(\"silhouette score\")\n",
    "\n",
    "ax2.plot(list(db.keys()), list(db.values()), color=\"red\", marker='o')\n",
    "ax2.set_xlabel(\"Number of cluster\")\n",
    "ax2.set_ylabel(\"Davies-Bouldin score\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27f65d20",
   "metadata": {},
   "source": [
    "## Kmeans et graphs TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3101852d",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "from sklearn.manifold import TSNE\n",
    "import matplotlib.cm as cm\n",
    "n_clusters=4\n",
    "\n",
    "sample_scaled = preprocessing_pipeline.fit_transform(sample)\n",
    "fig, ax2 = plt.subplots(1)\n",
    "fig.set_size_inches(10, 10)\n",
    "\n",
    "\n",
    "clusterer = KMeans(n_clusters=n_clusters, random_state=10, max_iter=400, n_init=50)\n",
    "cluster_labels = clusterer.fit_predict(sample_scaled)\n",
    "labels_scale = clusterer.labels_\n",
    "\n",
    "# The silhouette_score gives the average value for all the samples.\n",
    "# This gives a perspective into the density and separation of the formed clusters\n",
    "    \n",
    "silhouette_avg = silhouette_score(sample_scaled, cluster_labels)\n",
    "print(\"for n clusters =\", n_clusters, \"average silhouette_score is :\", silhouette_avg)\n",
    "\n",
    "# Compute the silhouette scores for each sample\n",
    "sample_silhouette_values = silhouette_samples(sample_scaled, cluster_labels)\n",
    "\n",
    "# 2nd Plot showing the actual clusters formed\n",
    "\n",
    "\n",
    "reducer1=TSNE(n_components=2)\n",
    "data_reduced1 = reducer1.fit_transform(sample_scaled)\n",
    "\n",
    "\n",
    "sns.scatterplot(data_reduced1[:,0], data_reduced1[:,1], palette='deep', hue=cluster_labels, legend=\"full\",alpha=0.7,ax=ax2)\n",
    "\n",
    "\n",
    "ax2.set_title(\"clustered data\")\n",
    "ax2.set_xlabel(\"feature space for 1st feature\")\n",
    "ax2.set_ylabel(\"feature space for 2nd feature\")\n",
    "plt.suptitle((\"TSNE for KMeans clustering\"\n",
    "                  \"with n clusters = %d\" % n_clusters),\n",
    "                 fontsize=14, fontweight='bold')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5ffdc750",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "# Définitiondu dataframe preprocessed\n",
    "sample_scaled = preprocessing_pipeline.fit_transform(sample)\n",
    "df_sample_scaled = pd.DataFrame(sample_scaled,\n",
    "                                 index=sample.index,\n",
    "                                 columns=sample.columns)\n",
    "\n",
    "\n",
    "n_clusters=4\n",
    "\n",
    "clusterer = KMeans(n_clusters=n_clusters, random_state=10, max_iter=400, n_init=50)\n",
    "cluster_labels = clusterer.fit_predict(sample_scaled)\n",
    "labels_scale = clusterer.labels_\n",
    "\n",
    "# Dataframe of identified clusters\n",
    "df_cluster_labels = pd.DataFrame(labels_scale, index = df_sample_scaled.index)\n",
    "df_cluster_labels.columns = ['cluster_labels']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8bb5c9a4",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Distribution of clients in clusters\n",
    "sample_clustered = df_sample_scaled.join(df_cluster_labels)\n",
    "sample_clustered['cluster_labels'].value_counts(normalize=True).sort_index().plot.bar(figsize=(9, 4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "205b03ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Distribution of clients in clusters\n",
    "sample_clustered['cluster_labels'].value_counts().sort_index()/len(sample_clustered)*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7077937a",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_test = sample_clustered.groupby(['cluster_labels']).agg({\n",
    "        'RecencyScore': 'size',\n",
    "        'FrequencyScore':'size',\n",
    "        'MonetaryScore':'size'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "781c8376",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_test.sort_index().plot.bar(figsize=(9, 4))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60d78862",
   "metadata": {},
   "source": [
    "## DBSCAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c2090895",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "from sklearn.cluster import DBSCAN\n",
    "from sklearn import metrics\n",
    "from sklearn.manifold import TSNE\n",
    "import matplotlib.cm as cm\n",
    "\n",
    "\n",
    "sample_scaled = preprocessing_pipeline.fit_transform(sample)\n",
    "fig, ax2 = plt.subplots(1)\n",
    "fig.set_size_inches(10, 10)\n",
    "\n",
    "\n",
    "# Compute DBSCAN\n",
    "db = DBSCAN(eps=0.05, min_samples=10).fit(sample_scaled)\n",
    "core_samples_mask = np.zeros_like(db.labels_, dtype=bool)\n",
    "core_samples_mask[db.core_sample_indices_] = True\n",
    "labels = db.labels_\n",
    "\n",
    "\n",
    "# 2nd Plot showing the actual clusters formed\n",
    "\n",
    "\n",
    "reducer1=TSNE(n_components=2)\n",
    "data_reduced1 = reducer1.fit_transform(sample_scaled)\n",
    "\n",
    "sns.scatterplot(data_reduced1[:,0], data_reduced1[:,1], palette='deep', hue=labels, legend=\"full\",alpha=0.7,ax=ax2)\n",
    "\n",
    "\n",
    "ax2.set_title(\"clustered data\")\n",
    "ax2.set_xlabel(\"feature space for 1st feature\")\n",
    "ax2.set_ylabel(\"feature space for 2nd feature\")\n",
    "plt.suptitle(\"TSNE for DBSCAN\", fontsize=14, fontweight='bold')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5cfe3dfc",
   "metadata": {},
   "source": [
    "# variables latentes des données et partitionnement avec un classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "701732ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Définitiondu dataframe preprocessed\n",
    "sample_imputed = preprocessing_pipeline.fit_transform(sample)\n",
    "df_sample_imputed = pd.DataFrame(sample_imputed,\n",
    "                                 index=sample.index,\n",
    "                                 columns=sample.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f882eecb",
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_scaled = preprocessing_pipeline.fit_transform(sample)\n",
    "sample_scaled = pd.DataFrame(sample_scaled,  index = sample.index, columns = sample.columns)\n",
    "n_clusters=4\n",
    "\n",
    "clusterer = KMeans(n_clusters=n_clusters, random_state=10, max_iter=400, n_init=50)\n",
    "cluster_labels = clusterer.fit_predict(sample_scaled)\n",
    "labels_scale = clusterer.labels_\n",
    "\n",
    "# Dataframe of identified clusters\n",
    "df_cluster_labels = pd.DataFrame(cluster_labels, index = sample_scaled.index)\n",
    "df_cluster_labels.columns = ['cluster_labels']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e24707f4",
   "metadata": {},
   "source": [
    "## clustering hiérarchique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9b559ca0",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# hierachical clustering for flexibiliy and small ammount of features (<50)\n",
    "# random forest for feature importance on a classification problem \n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from matplotlib import pyplot\n",
    "\n",
    "X, y = df_sample_imputed, df_cluster_labels\n",
    "# define the model\n",
    "classeur = RandomForestClassifier()\n",
    "# fit the model\n",
    "classeur.fit(X, y)\n",
    "# get importance\n",
    "importance = classeur.feature_importances_\n",
    "df_importance = pd.DataFrame(importance,\n",
    "                             index=X.columns,\n",
    "                             )\n",
    "df_importance = df_importance.rename(columns={0: 'importance'})\n",
    "df_importance = df_importance.sort_values(by = ['importance'], ascending=False)\n",
    "# plot feature importance\n",
    " # .nlargest(4) # pandas Series method which will return a subset of the series with the largest n values.\n",
    "\n",
    "sns.barplot(x='importance',\n",
    "            y=df_importance.index, data=df_importance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "23027fd4",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Distribution of clients in clusters\n",
    "sample_clustered['cluster_labels'].value_counts().sort_index()/len(sample_clustered)*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ff6b7be6",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Parallel Coordinates Plots pour la caractérisation des clusters\n",
    "\n",
    "import plotly.graph_objects as go\n",
    "\n",
    "fig = go.Figure(data=\n",
    "    go.Parcoords(\n",
    "        line = dict(color = sample_clustered['cluster_labels'],\n",
    "                   colorscale = [[0,'red'],[0.1,'blue'],[0.2,'yellow'], [0.3,'green'],[0.4,'darkgreen'],\n",
    "                                 [0.5,'orange'], [0.6,'purple'],[0.7,'lightblue'],[1,'gold']]),\n",
    "        \n",
    "        dimensions = list([\n",
    "            dict(range = [sample_clustered['RecencyScore'].min(),sample_clustered['RecencyScore'].max()],\n",
    "                 #constraintrange = [100000,150000],\n",
    "                 label = \"RecencyScore\", values = sample_clustered['RecencyScore']),\n",
    "            \n",
    "            dict(range = [-1,6],\n",
    "                 tickvals = [0,1,2,3,4,5],\n",
    "                 label = 'cluster_labels', values = sample_clustered['cluster_labels']),\n",
    "            \n",
    "            dict(range = [sample_clustered['FrequencyScore'].min(),sample_clustered['FrequencyScore'].max()],\n",
    "                 #constraintrange = [100000,150000],\n",
    "                 label = \"FrequencyScore\", values = sample_clustered['FrequencyScore']),\n",
    "            \n",
    "            dict(range = [sample_clustered['MonetaryScore'].min(),sample_clustered['MonetaryScore'].max()],\n",
    "                 #constraintrange = [100000,150000],\n",
    "                 label = \"MonetaryScore\", values = sample_clustered['MonetaryScore'])])  \n",
    "    )\n",
    ")\n",
    "\n",
    "fig.update_layout(\n",
    "    plot_bgcolor = 'white',\n",
    "    paper_bgcolor = 'white'\n",
    ")\n",
    "\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "531200b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Valeurs moyennes par variables et par clusters\n",
    "column_names = list(sample_clustered.columns)\n",
    "\n",
    "df_mean = pd.DataFrame(columns = column_names)\n",
    "\n",
    "for i in range(4):\n",
    "    df_mean = df_mean.append((pd.DataFrame(sample_clustered[sample_clustered['cluster_labels']==i].mean())).transpose())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "434206f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "046b8bef",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df_mean.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6cb5d246",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean['MonetaryScore'].min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a40d30a4",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Parallel Coordinates Plots pour la caractérisation des clusters\n",
    "\n",
    "import plotly.graph_objects as go\n",
    "\n",
    "fig = go.Figure(data=\n",
    "    go.Parcoords(\n",
    "        line = dict(color = df_mean['cluster_labels'],\n",
    "                   colorscale = [[0,'red'],[0.1,'blue'],[0.2,'yellow'], [0.3,'green'],[0.4,'darkgreen'],\n",
    "                                 [0.5,'orange'], [0.6,'purple'],[0.7,'lightblue'],[1,'gold']]),\n",
    "        \n",
    "        dimensions = list([\n",
    "            \n",
    "            dict(range = [0,4],\n",
    "                 tickvals = [0,1,2,3],\n",
    "                 label = 'cluster_labels', values = df_mean['cluster_labels']),\n",
    "            \n",
    "            dict(range = [sample_clustered['MonetaryScore'].min(),sample_clustered['MonetaryScore'].max()],\n",
    "                 label = \"MonetaryScore\", values = sample_clustered['MonetaryScore'].astype('int64')),\n",
    "            \n",
    "            dict(range = [df_mean['FrequencyScore'].min(),df_mean['FrequencyScore'].max()],\n",
    "                 label = \"FrequencyScore\", values = df_mean['FrequencyScore'].astype('int64')),\n",
    "            \n",
    "            dict(range = [sample_clustered['RecencyScore'].min(),sample_clustered['RecencyScore'].max()],\n",
    "                 label = \"RecencyScore\", values = sample_clustered['RecencyScore'].astype('int64'))])\n",
    "\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.update_layout(\n",
    "    plot_bgcolor = 'white',\n",
    "    paper_bgcolor = 'white'\n",
    ")\n",
    "\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "76235110",
   "metadata": {},
   "source": [
    "## Stabilité du Kmeans dans le temps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78f9673e",
   "metadata": {},
   "outputs": [],
   "source": [
    "rfm.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9cb40a46",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "rfm.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "394894a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_scaled = preprocessing_pipeline.fit_transform(rfm)\n",
    "\n",
    "# transformation into dataframe\n",
    "df_customer_dataset_reduced = pd.DataFrame(sample_scaled, \n",
    "                                    index = rfm.index, \n",
    "                                    columns = rfm.columns)\n",
    "\n",
    "df_customer_dataset_reduced_time = pd.merge(model[['purchase_year', 'purchase_month']], \n",
    "                                            df_customer_dataset_reduced, \n",
    "                                            left_index=True, right_index=True,  how = 'outer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d5c5191b",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_clusters=4\n",
    "clusterer = KMeans(n_clusters=n_clusters, random_state=10, max_iter=400, n_init=50)\n",
    "cluster_labels = clusterer.fit_predict(sample_scaled)\n",
    "labels_scale = clusterer.labels_\n",
    "\n",
    "cluster_labels = clusterer.fit_predict(sample_scaled)\n",
    "df_cluster_labels = pd.DataFrame(cluster_labels, index = df_customer_dataset_reduced.index)\n",
    "df_cluster_labels.rename(columns={0: \"clusters\"}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "44cd38f6",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics.cluster import adjusted_rand_score\n",
    "# list of purchase_year and purchase_month\n",
    "year =  (df_customer_dataset_reduced_time.purchase_year.astype(int).unique()).tolist() \n",
    "month =  (df_customer_dataset_reduced_time.purchase_month.astype(int).unique()).tolist()\n",
    "\n",
    "datasets = []\n",
    "ari = []\n",
    "\n",
    "# for each month of each year we extract the corresponding part of the dataset and we store it in a list of monthly datasets\n",
    "for y in year:\n",
    "    for m in month:\n",
    "        data = df_customer_dataset_reduced_time.loc[(df_customer_dataset_reduced_time.purchase_year == y)\n",
    "                                                    & (df_customer_dataset_reduced_time.purchase_month == m),:]\n",
    "        datasets.append(data)\n",
    "        \n",
    "df_merge = []\n",
    "\n",
    "# for each element of the monthly dataset list, calculate the ARI score between the labels of a monthly dataset \n",
    "# and those of the global dataset for the same period\n",
    "\n",
    "for k in range(len(datasets)):\n",
    "    df_merge.append(datasets[k])\n",
    "    if len(df_merge) == 1 :\n",
    "        df_final = df_merge[0]\n",
    "        if len(df_final) >= 4:\n",
    "            df_final.drop(['purchase_year', 'purchase_month'], axis = 1 , inplace = True)\n",
    "            clusterer.fit(df_final)\n",
    "            labels_true = df_cluster_labels.loc[df_final.index, 'clusters'].tolist()\n",
    "            ari.append(adjusted_rand_score(labels_true, clusterer.labels_))\n",
    "    else:\n",
    "        df_final = pd.concat(df_merge)\n",
    "        if len(df_final) >= 4:\n",
    "            df_final.drop(['purchase_year', 'purchase_month'], axis = 1 , inplace = True)\n",
    "            clusterer.fit(df_final)\n",
    "            labels_true = df_cluster_labels.loc[df_final.index, 'clusters'].tolist()\n",
    "            ari.append(adjusted_rand_score(labels_true, clusterer.labels_))\n",
    "\n",
    "ari.reverse()\n",
    "\n",
    "#plot of results\n",
    "fig = plt.figure(figsize=(16, 4))\n",
    "plt.plot(range(len(ari)), ari, marker='o')\n",
    "plt.xlabel('nombre de mois')\n",
    "plt.ylabel('indice ARI')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}