Clustering_final.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn import preprocessing\n",
    "from scipy import stats\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.cluster import KMeans\n",
    "import matplotlib.pyplot as plt\n",
    "import prince\n",
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "import seaborn as sns\n",
    "import chart_studio.plotly as py\n",
    "import plotly.graph_objs as go\n",
    "from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preparation for the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 444,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', 500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 445,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('/Users/naitochieri/Desktop/Hackathon/Datasets/combined.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 446,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Unnamed: 0', 'index', 'Unnamed: 0.1', 'Unnamed: 0.1.1', 'account_id_x',\n",
       "       'vertical', 'industry', 'client_country', 'unique_client',\n",
       "       'client_continent', 'client_region', 'OS', 'device', 'environment',\n",
       "       'criteo_product', 'month', 'time', 'year', 'spend_criteo_euro',\n",
       "       'sales_criteo', 'CPO', 'revenue_criteo_euro', 'AOV', 'COS',\n",
       "       'churn_date', 'churn', 'start_time', 'Android', 'MacOS', 'Other',\n",
       "       'Unknown_x', 'iOs', 'Desktop', 'Mobile', 'Unknown_y', 'app', 'web',\n",
       "       'awareness', 'consideration', 'conversion', 'account_id_y', 'products',\n",
       "       'account_id', 'onsite_sales', 'onsite_revenue_euro'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 446,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 447,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df[['unique_client','month', 'time', 'vertical', 'industry', 'client_country',\n",
    "       'client_continent', 'client_region', 'OS', 'device', 'environment',\n",
    "       'criteo_product', 'spend_criteo_euro', 'sales_criteo', 'CPO',\n",
    "       'revenue_criteo_euro', 'AOV', 'COS', 'Android', 'MacOS', 'Other', 'iOs',\n",
    "       'Desktop', 'Mobile', 'app', 'web', 'awareness', 'consideration',\n",
    "       'conversion', 'products', 'onsite_sales', 'onsite_revenue_euro','churn_date']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 448,
   "metadata": {},
   "outputs": [],
   "source": [
    "df2 = df[['vertical', 'industry', 'client_country',\n",
    "       'client_continent', 'client_region', 'OS', 'device',\n",
    "       'environment', 'criteo_product', 'month', 'time',\n",
    "         'spend_criteo_euro', 'sales_criteo', 'CPO', 'revenue_criteo_euro',\n",
    "       'AOV', 'COS', 'churn_date']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 449,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(X.shape[0]):\n",
    "    if X.loc[i,'churn_date'] == 0:\n",
    "        X.loc[i,'churn_date'] = int(X['time'][i]) + 5.01\n",
    "        \n",
    "for i in range(df2.shape[0]):\n",
    "    if df2.loc[i,'churn_date'] == 0:\n",
    "        df2.loc[i,'churn_date'] = int(df2['time'][i]) + 5.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 450,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_cols = ['month', 'time',\n",
    "         'spend_criteo_euro', 'sales_criteo', 'CPO', 'revenue_criteo_euro',\n",
    "       'AOV', 'COS']\n",
    "cate_cols = df2.columns.drop('churn_date').drop(num_cols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 451,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:3494: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n"
     ]
    }
   ],
   "source": [
    "X[num_cols] = X[num_cols].apply(pd.to_numeric, errors='coerce')\n",
    "X['churn_date'] = X['churn_date'].apply(pd.to_numeric, errors='coerce')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 452,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n"
     ]
    }
   ],
   "source": [
    "df2[num_cols] = df2[num_cols].apply(pd.to_numeric, errors='coerce')\n",
    "df2['churn_date'] = df2['churn_date'].apply(pd.to_numeric, errors='coerce')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 453,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.fillna(0)\n",
    "df2 = df2.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 454,
   "metadata": {},
   "outputs": [],
   "source": [
    "eps = 1e-15\n",
    "for i in range(X.shape[0]):\n",
    "    number_dec_str = str(X['churn_date'][i]-int(X['churn_date'][i]))[1:]\n",
    "    number_dec = float(number_dec_str)\n",
    "    for j in range(1,13):\n",
    "            if abs(number_dec - 0.01*j) < eps:\n",
    "                X.loc[i,'churn_date'] = int(X['churn_date'][i]) + j/13\n",
    "\n",
    "eps = 1e-15\n",
    "for i in range(df2.shape[0]):\n",
    "    number_dec_str = str(df2['churn_date'][i]-int(df2['churn_date'][i]))[1:]\n",
    "    number_dec = float(number_dec_str)\n",
    "    for j in range(1,13):\n",
    "            if abs(number_dec - 0.01*j) < eps:\n",
    "                df2.loc[i,'churn_date'] = int(df2['churn_date'][i]) + j/13"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 455,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "vertical                object\n",
       "industry                object\n",
       "client_country          object\n",
       "client_continent        object\n",
       "client_region           object\n",
       "OS                      object\n",
       "device                  object\n",
       "environment             object\n",
       "criteo_product          object\n",
       "month                    int64\n",
       "time                   float64\n",
       "spend_criteo_euro        int64\n",
       "sales_criteo             int64\n",
       "CPO                    float64\n",
       "revenue_criteo_euro      int64\n",
       "AOV                    float64\n",
       "COS                      int64\n",
       "churn_date             float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 455,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X2 = df2\n",
    "X2.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 456,
   "metadata": {},
   "outputs": [],
   "source": [
    "X['country_industry'] = X['client_country'] + '_' + X['industry']\n",
    "X2['country_industry'] = X2['client_country'] + '_' + X2['industry']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predict churn date to calculate tenure using our trained Xgboost regression model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:33:12] WARNING: src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "import pickle\n",
    "filename = 'churn_date_model.sav'\n",
    "model = pickle.load(open(filename, 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_pred(X):\n",
    "    X3 = X.drop(columns=['churn_date'])\n",
    "    y = X['churn_date']\n",
    "    encoder = preprocessing.LabelEncoder()\n",
    "    X3['vertical'] = encoder.fit_transform(X['vertical'])\n",
    "    X3['industry'] = encoder.fit_transform(X['industry'])\n",
    "    X3['client_country'] = encoder.fit_transform(X['client_country'])\n",
    "    X3['client_continent'] = encoder.fit_transform(X['client_continent'])\n",
    "    X3['client_region'] = encoder.fit_transform(X['client_region'])\n",
    "    X3['OS'] = encoder.fit_transform(X['OS'])\n",
    "    X3['device'] = encoder.fit_transform(X['device'])\n",
    "    X3['environment'] = encoder.fit_transform(X['environment'])\n",
    "    X3['criteo_product'] = encoder.fit_transform(X['criteo_product'])\n",
    "    #X3['index'] = \"\"\n",
    "    #for i, j in X3.iterrows():\n",
    "       # X3.loc[i, 'index'] = i\n",
    "    cols = ['month', 'time', 'vertical', 'industry', 'client_country', 'client_continent', 'client_region', 'OS', 'device', 'environment', 'criteo_product', 'spend_criteo_euro', 'sales_criteo', 'CPO', 'revenue_criteo_euro', 'AOV', 'COS', 'Android', 'MacOS', 'Other', 'iOs', 'Desktop', 'Mobile', 'app', 'web', 'awareness', 'consideration', 'conversion', 'products', 'onsite_sales', 'onsite_revenue_euro']\n",
    "    X3 = X3[cols]\n",
    "    \n",
    "    return model.predict(X3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 331,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Identify the correct month given a churn date\n",
    "def identify_month(churn_date):\n",
    "    eps = 1e-15\n",
    "    number_dec_str = str(churn_date-int(churn_date))[1:]\n",
    "    number_dec = float(number_dec_str)\n",
    "    entry = False\n",
    "    for j in range(1,13):\n",
    "        if (number_dec < j/13.0 or abs(number_dec - j/13) < eps) and number_dec > (j-1)/13.0:\n",
    "            entry = True\n",
    "            return j\n",
    "    if not entry:\n",
    "        return 12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Identify the churn dates in the format (year, month) for all good clients in each cluster\n",
    "def get_final_date(pred_list):\n",
    "    final_dates = []\n",
    "    for i in range(len(pred_list)):\n",
    "        final_dates.append((int(pred_list[i]), identify_month(pred_list[i])))\n",
    "    return final_dates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = get_pred(X)\n",
    "churn_date_pred = get_final_date(pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>2022</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>2020</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>2022</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>2024</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>2021</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16245</td>\n",
       "      <td>2022</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16246</td>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16247</td>\n",
       "      <td>2022</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16248</td>\n",
       "      <td>2022</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16249</td>\n",
       "      <td>2022</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>16250 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          0   1\n",
       "0      2022   4\n",
       "1      2020   3\n",
       "2      2022   4\n",
       "3      2024   8\n",
       "4      2021   5\n",
       "...     ...  ..\n",
       "16245  2022   6\n",
       "16246  2023   1\n",
       "16247  2022  12\n",
       "16248  2022   4\n",
       "16249  2022   9\n",
       "\n",
       "[16250 rows x 2 columns]"
      ]
     },
     "execution_count": 343,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_date_pred = pd.DataFrame(data=churn_date_pred, index=X.index)\n",
    "churn_date_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 16250 entries, 0 to 16249\n",
      "Data columns (total 2 columns):\n",
      "0    16250 non-null int64\n",
      "1    16250 non-null int64\n",
      "dtypes: int64(2)\n",
      "memory usage: 254.0 KB\n"
     ]
    }
   ],
   "source": [
    "churn_date_pred.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 346,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 4,  3,  8,  5,  7,  6, 10, 11,  2,  9, 12,  1])"
      ]
     },
     "execution_count": 346,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_date_pred[1].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "metadata": {},
   "outputs": [],
   "source": [
    "churn_date_pred[1] = churn_date_pred[1]*0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        2022.04\n",
       "1        2020.03\n",
       "2        2022.04\n",
       "3        2024.08\n",
       "4        2021.05\n",
       "          ...   \n",
       "16245    2022.06\n",
       "16246    2023.01\n",
       "16247    2022.12\n",
       "16248    2022.04\n",
       "16249    2022.09\n",
       "Length: 16250, dtype: float64"
      ]
     },
     "execution_count": 349,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_date_pred = churn_date_pred[0] +  churn_date_pred[1]\n",
    "churn_date_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 351,
   "metadata": {},
   "outputs": [],
   "source": [
    "X['churn_date_pred'] = churn_date_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 418,
   "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>unique_client</th>\n",
       "      <th>month</th>\n",
       "      <th>time</th>\n",
       "      <th>vertical</th>\n",
       "      <th>industry</th>\n",
       "      <th>client_country</th>\n",
       "      <th>client_continent</th>\n",
       "      <th>client_region</th>\n",
       "      <th>OS</th>\n",
       "      <th>device</th>\n",
       "      <th>environment</th>\n",
       "      <th>criteo_product</th>\n",
       "      <th>spend_criteo_euro</th>\n",
       "      <th>sales_criteo</th>\n",
       "      <th>CPO</th>\n",
       "      <th>revenue_criteo_euro</th>\n",
       "      <th>AOV</th>\n",
       "      <th>COS</th>\n",
       "      <th>Android</th>\n",
       "      <th>MacOS</th>\n",
       "      <th>Other</th>\n",
       "      <th>iOs</th>\n",
       "      <th>Desktop</th>\n",
       "      <th>Mobile</th>\n",
       "      <th>app</th>\n",
       "      <th>web</th>\n",
       "      <th>awareness</th>\n",
       "      <th>consideration</th>\n",
       "      <th>conversion</th>\n",
       "      <th>products</th>\n",
       "      <th>onsite_sales</th>\n",
       "      <th>onsite_revenue_euro</th>\n",
       "      <th>churn_date</th>\n",
       "      <th>country_industry</th>\n",
       "      <th>churn_date_pred</th>\n",
       "      <th>duration</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>ga_bc66d247-ae5d-e311-a5f9-005056bb7e02_RUSSIA...</td>\n",
       "      <td>3</td>\n",
       "      <td>2017.03</td>\n",
       "      <td>RETAIL</td>\n",
       "      <td>Electronics / High Tech</td>\n",
       "      <td>RUSSIAN FEDERATION</td>\n",
       "      <td>EMEA</td>\n",
       "      <td>RUSSIA</td>\n",
       "      <td>Android</td>\n",
       "      <td>Mobile</td>\n",
       "      <td>web</td>\n",
       "      <td>conversion</td>\n",
       "      <td>1030</td>\n",
       "      <td>168</td>\n",
       "      <td>6.0</td>\n",
       "      <td>22929</td>\n",
       "      <td>136.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1030.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1030.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1030.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1030.0</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>3093.0</td>\n",
       "      <td>416103.72</td>\n",
       "      <td>2018.06</td>\n",
       "      <td>RUSSIAN FEDERATION_Electronics / High Tech</td>\n",
       "      <td>2022.04</td>\n",
       "      <td>5.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>ga_bc784dcb-17fe-e311-a3df-005056bb7e02_CANADA</td>\n",
       "      <td>4</td>\n",
       "      <td>2017.04</td>\n",
       "      <td>RETAIL</td>\n",
       "      <td>Gifts &amp; Flowers</td>\n",
       "      <td>CANADA</td>\n",
       "      <td>AMERICAS</td>\n",
       "      <td>CANADA</td>\n",
       "      <td>Android</td>\n",
       "      <td>Mobile</td>\n",
       "      <td>web</td>\n",
       "      <td>conversion</td>\n",
       "      <td>1966</td>\n",
       "      <td>4</td>\n",
       "      <td>492.0</td>\n",
       "      <td>246</td>\n",
       "      <td>61.0</td>\n",
       "      <td>8</td>\n",
       "      <td>1966.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1966.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1966.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1966.0</td>\n",
       "      <td>4604.0</td>\n",
       "      <td>585.0</td>\n",
       "      <td>183438.72</td>\n",
       "      <td>2018.09</td>\n",
       "      <td>CANADA_Gifts &amp; Flowers</td>\n",
       "      <td>2020.03</td>\n",
       "      <td>2.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>ga_bc784dcb-17fe-e311-a3df-005056bb7e02_UNITED...</td>\n",
       "      <td>9</td>\n",
       "      <td>2017.09</td>\n",
       "      <td>RETAIL</td>\n",
       "      <td>Gifts &amp; Flowers</td>\n",
       "      <td>UNITED KINGDOM</td>\n",
       "      <td>EMEA</td>\n",
       "      <td>UK</td>\n",
       "      <td>Android</td>\n",
       "      <td>Mobile</td>\n",
       "      <td>web</td>\n",
       "      <td>conversion</td>\n",
       "      <td>1377</td>\n",
       "      <td>168</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8606</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1377.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1377.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1377.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1377.0</td>\n",
       "      <td>13958.0</td>\n",
       "      <td>21107.0</td>\n",
       "      <td>3858810.09</td>\n",
       "      <td>2018.05</td>\n",
       "      <td>UNITED KINGDOM_Gifts &amp; Flowers</td>\n",
       "      <td>2022.04</td>\n",
       "      <td>4.95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>ga_bc784dcb-17fe-e311-a3df-005056bb7e02_UNITED...</td>\n",
       "      <td>7</td>\n",
       "      <td>2019.07</td>\n",
       "      <td>RETAIL</td>\n",
       "      <td>Gifts &amp; Flowers</td>\n",
       "      <td>UNITED STATES</td>\n",
       "      <td>AMERICAS</td>\n",
       "      <td>US</td>\n",
       "      <td>Android</td>\n",
       "      <td>Desktop</td>\n",
       "      <td>web</td>\n",
       "      <td>conversion</td>\n",
       "      <td>82</td>\n",
       "      <td>4</td>\n",
       "      <td>21.0</td>\n",
       "      <td>72</td>\n",
       "      <td>18.0</td>\n",
       "      <td>1</td>\n",
       "      <td>82.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>4990.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>11705.52</td>\n",
       "      <td>0.00</td>\n",
       "      <td>UNITED STATES_Gifts &amp; Flowers</td>\n",
       "      <td>2024.08</td>\n",
       "      <td>5.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>ga_8fff21b1-4fad-e511-9606-005056bb2344_RUSSIA...</td>\n",
       "      <td>6</td>\n",
       "      <td>2017.06</td>\n",
       "      <td>RETAIL</td>\n",
       "      <td>Accessories / Jewelry / Luxury Goods</td>\n",
       "      <td>RUSSIAN FEDERATION</td>\n",
       "      <td>EMEA</td>\n",
       "      <td>RUSSIA</td>\n",
       "      <td>Android</td>\n",
       "      <td>Mobile</td>\n",
       "      <td>web</td>\n",
       "      <td>conversion</td>\n",
       "      <td>74</td>\n",
       "      <td>20</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1174</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>5147.0</td>\n",
       "      <td>686.0</td>\n",
       "      <td>87426.68</td>\n",
       "      <td>2018.02</td>\n",
       "      <td>RUSSIAN FEDERATION_Accessories / Jewelry / Lux...</td>\n",
       "      <td>2021.05</td>\n",
       "      <td>3.99</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                       unique_client  month     time vertical  \\\n",
       "0  ga_bc66d247-ae5d-e311-a5f9-005056bb7e02_RUSSIA...      3  2017.03   RETAIL   \n",
       "1     ga_bc784dcb-17fe-e311-a3df-005056bb7e02_CANADA      4  2017.04   RETAIL   \n",
       "2  ga_bc784dcb-17fe-e311-a3df-005056bb7e02_UNITED...      9  2017.09   RETAIL   \n",
       "3  ga_bc784dcb-17fe-e311-a3df-005056bb7e02_UNITED...      7  2019.07   RETAIL   \n",
       "4  ga_8fff21b1-4fad-e511-9606-005056bb2344_RUSSIA...      6  2017.06   RETAIL   \n",
       "\n",
       "                               industry      client_country client_continent  \\\n",
       "0               Electronics / High Tech  RUSSIAN FEDERATION             EMEA   \n",
       "1                       Gifts & Flowers              CANADA         AMERICAS   \n",
       "2                       Gifts & Flowers      UNITED KINGDOM             EMEA   \n",
       "3                       Gifts & Flowers       UNITED STATES         AMERICAS   \n",
       "4  Accessories / Jewelry / Luxury Goods  RUSSIAN FEDERATION             EMEA   \n",
       "\n",
       "  client_region       OS   device environment criteo_product  \\\n",
       "0        RUSSIA  Android   Mobile         web     conversion   \n",
       "1        CANADA  Android   Mobile         web     conversion   \n",
       "2            UK  Android   Mobile         web     conversion   \n",
       "3            US  Android  Desktop         web     conversion   \n",
       "4        RUSSIA  Android   Mobile         web     conversion   \n",
       "\n",
       "   spend_criteo_euro  sales_criteo    CPO  revenue_criteo_euro    AOV  COS  \\\n",
       "0               1030           168    6.0                22929  136.0    0   \n",
       "1               1966             4  492.0                  246   61.0    8   \n",
       "2               1377           168    8.0                 8606   51.0    0   \n",
       "3                 82             4   21.0                   72   18.0    1   \n",
       "4                 74            20    4.0                 1174   59.0    0   \n",
       "\n",
       "   Android  MacOS  Other  iOs  Desktop  Mobile  app     web  awareness  \\\n",
       "0   1030.0    0.0    0.0  0.0      0.0  1030.0  0.0  1030.0        0.0   \n",
       "1   1966.0    0.0    0.0  0.0      0.0  1966.0  0.0  1966.0        0.0   \n",
       "2   1377.0    0.0    0.0  0.0      0.0  1377.0  0.0  1377.0        0.0   \n",
       "3     82.0    0.0    0.0  0.0     82.0     0.0  0.0    82.0        0.0   \n",
       "4     74.0    0.0    0.0  0.0      0.0    74.0  0.0    74.0        0.0   \n",
       "\n",
       "   consideration  conversion  products  onsite_sales  onsite_revenue_euro  \\\n",
       "0            0.0      1030.0    4500.0        3093.0            416103.72   \n",
       "1            0.0      1966.0    4604.0         585.0            183438.72   \n",
       "2            0.0      1377.0   13958.0       21107.0           3858810.09   \n",
       "3            0.0        82.0    4990.0          82.0             11705.52   \n",
       "4            0.0        74.0    5147.0         686.0             87426.68   \n",
       "\n",
       "   churn_date                                   country_industry  \\\n",
       "0     2018.06         RUSSIAN FEDERATION_Electronics / High Tech   \n",
       "1     2018.09                             CANADA_Gifts & Flowers   \n",
       "2     2018.05                     UNITED KINGDOM_Gifts & Flowers   \n",
       "3        0.00                      UNITED STATES_Gifts & Flowers   \n",
       "4     2018.02  RUSSIAN FEDERATION_Accessories / Jewelry / Lux...   \n",
       "\n",
       "   churn_date_pred  duration  \n",
       "0          2022.04      5.01  \n",
       "1          2020.03      2.99  \n",
       "2          2022.04      4.95  \n",
       "3          2024.08      5.01  \n",
       "4          2021.05      3.99  "
      ]
     },
     "execution_count": 418,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 354,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(X.shape[0]):\n",
    "    X.loc[i,'duration'] = float(X['churn_date_pred'][i]) - X['time'][i]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dimensionality Reduction with FAMD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "famd = prince.FAMD(n_components=2,n_iter=3,copy=True,check_input=True,engine='auto',random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/prince/one_hot.py:35: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:257: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:3456: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/ops/__init__.py:1641: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:339: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py:6289: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py:5884: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/prince/one_hot.py:35: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:257: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:3456: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/ops/__init__.py:1641: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:339: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py:392: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "famd = prince.FAMD(n_components=2,n_iter=3,copy=True,check_input=True,engine='auto',random_state=42)\n",
    "famd = famd.fit(X2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/prince/one_hot.py:35: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:257: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:3456: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/ops/__init__.py:1641: FutureWarning:\n",
      "\n",
      "SparseSeries is deprecated and will be removed in a future version.\n",
      "Use a Series with sparse values instead.\n",
      "\n",
      "    >>> series = pd.Series(pd.SparseArray(...))\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/sparse/frame.py:339: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n",
      "/Users/naitochieri/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py:392: FutureWarning:\n",
      "\n",
      "SparseDataFrame is deprecated and will be removed in a future version.\n",
      "Use a regular DataFrame whose columns are SparseArrays instead.\n",
      "\n",
      "See http://pandas.pydata.org/pandas-docs/stable/user_guide/sparse.html#migrating for more.\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "X2d = famd.row_coordinates(X2)\n",
    "X2d_arr = X2d.to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 355,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_labels_col(X2, column_name):\n",
    "    labels = []\n",
    "    unique_values = list(df[column_name].unique())\n",
    "    for i in range(len(df[column_name])):\n",
    "        for j in range(len(unique_values)):\n",
    "            if X2[column_name][i] == unique_values[j]:\n",
    "                labels.append(j)\n",
    "    return labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels_industry = get_labels_col(X, 'industry')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 357,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels_country = get_labels_col(X, 'client_country')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot the distances for country and industry"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 358,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x130a1d610>"
      ]
     },
     "execution_count": 358,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 7))\n",
    "plt.scatter(X2d[0], X2d[1], c = np.asarray(labels_industry).astype(float),s=70, alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 359,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x130a1c390>"
      ]
     },
     "execution_count": 359,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,7))\n",
    "plt.scatter(X2d[0], X2d[1], c= np.asarray(labels_country).astype(float), s=70, alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 360,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dim_reduction(X2):\n",
    "    famd = prince.FAMD(n_components=2,n_iter=3,copy=True,check_input=True,engine='auto',random_state=42)\n",
    "    famd = famd.fit(X2)\n",
    "    X2d = famd.row_coordinates(X2)\n",
    "    return X2d"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create function to find the best number of cluster to maximize the silhouette score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 361,
   "metadata": {},
   "outputs": [],
   "source": [
    "def maximize_cluster_score(X2d, low_value, high_value):\n",
    "    max_score = -1\n",
    "    best_cluster_number = 0\n",
    "    for i in range(low_value, high_value):\n",
    "        kmeans = KMeans(n_clusters=i).fit(X2d)\n",
    "        if silhouette_score(X2d, kmeans.labels_) > max_score:\n",
    "            max_score = silhouette_score(X2d, kmeans.labels_)\n",
    "            best_cluster_number = i\n",
    "    return best_cluster_number, max_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create function to run the model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_clusters_per_column(X2, X2d, class_name, cluster_number):\n",
    "    X2d_column = X2d[X2['country_industry'] == class_name]\n",
    "    kmeans = KMeans(n_clusters=cluster_number).fit(X2d_column)\n",
    "    y_kmeans = kmeans.predict(X2d_column)\n",
    "    plt.figure(figsize=(10,7))\n",
    "    plt.scatter(X2d_column[0], X2d_column[1], c= y_kmeans, s=800, cmap='viridis')\n",
    "    plt.title(class_name)\n",
    "    plt.xlabel(\"x\")\n",
    "    plt.ylabel(\"y\")\n",
    "    return kmeans.labels_, silhouette_score(X2d_column, kmeans.labels_), kmeans.cluster_centers_, X2d_column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 363,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "16250"
      ]
     },
     "execution_count": 363,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(X2['country_industry'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "544"
      ]
     },
     "execution_count": 364,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(X2['country_industry'] == 'FRANCE_Apparel')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 365,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "407"
      ]
     },
     "execution_count": 365,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(X2['country_industry'] == 'JAPAN_Apparel')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1224"
      ]
     },
     "execution_count": 366,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(X2['country_industry'] == 'UNITED STATES_Apparel')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clustering for 3 chosen Country x Industry"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_FRANCE_Apparel = X[X['country_industry'] == 'FRANCE_Apparel']\n",
    "df_FRANCE_Apparel_copy = X[X['country_industry'] == 'FRANCE_Apparel']\n",
    "df_FRANCE_Apparel = df_FRANCE_Apparel.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_JAPAN_Apparel = X[X['country_industry'] == 'JAPAN_Apparel']\n",
    "df_JAPAN_Apparel_copy = X[X['country_industry'] == 'JAPAN_Apparel']\n",
    "df_JAPAN_Apparel = df_JAPAN_Apparel.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_UNITED_STATES_Apparel = X[X['country_industry'] == 'UNITED STATES_Apparel']\n",
    "df_UNITED_STATES_Apparel_copy = X[X['country_industry'] == 'UNITED STATES_Apparel']\n",
    "df_UNITED_STATES_Apparel = df_UNITED_STATES_Apparel.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_FRANCE_Apparel['labels'], s1, c1, X2d_france = get_clusters_per_column(X, X2d, 'FRANCE_Apparel', 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_JAPAN_Apparel['labels'], s2, c2, X2d_japan = get_clusters_per_column(X, X2d, 'JAPAN_Apparel', 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 372,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_UNITED_STATES_Apparel['labels'], s3, c3, X2d_us = get_clusters_per_column(X, X2d, 'UNITED STATES_Apparel', 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_FRANCE_Apparel.to_csv('FRANCE_Apparel.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 374,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_JAPAN_Apparel.to_csv('JAPAN_Apparel.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 375,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_UNITED_STATES_Apparel.to_csv('UNITED_STATES_Apparel.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# We got very high silhouette score !"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 419,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9902152699203981 0.9907748642383635 0.989359765960901\n"
     ]
    }
   ],
   "source": [
    "print(s1,s2,s3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 420,
   "metadata": {},
   "outputs": [],
   "source": [
    "ind_france = list((X[X['country_industry'] == 'FRANCE_Apparel']).index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 421,
   "metadata": {},
   "outputs": [],
   "source": [
    "ind_japan = list((X[X['country_industry'] == 'JAPAN_Apparel']).index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 422,
   "metadata": {},
   "outputs": [],
   "source": [
    "ind_us = list((X[X['country_industry'] == 'UNITED STATES_Apparel']).index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 423,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Map new indices to the old ones\n",
    "def get_dict(ind, length):\n",
    "    new_dict = dict()\n",
    "    for i in range(length):\n",
    "        new_dict[i] = ind[i]\n",
    "    return new_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 424,
   "metadata": {},
   "outputs": [],
   "source": [
    "dict_france = get_dict(ind_france, df_FRANCE_Apparel.shape[0])\n",
    "dict_japan = get_dict(ind_japan, df_JAPAN_Apparel.shape[0])\n",
    "dict_us = get_dict(ind_us, df_UNITED_STATES_Apparel.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get average clients in each cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 425,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_average_clients(df, X2d, centroids):\n",
    "    row_labels = []\n",
    "    labels = list(df['labels'])\n",
    "    for label in set(labels):\n",
    "        row_list = []\n",
    "        for i in range(len(X2d)):\n",
    "            if labels[i] == label:\n",
    "                row_list.append(i)\n",
    "        row_labels.append(row_list)\n",
    "    \n",
    "    average_clients = []\n",
    "    for i in range(len(centroids)):\n",
    "        min_dist = -1\n",
    "        ind_min = 0\n",
    "        for j in range(len(row_labels[i])):\n",
    "            if min_dist == -1 or (X2d[row_labels[i][j]][0] - centroids[i][0])*(X2d[row_labels[i][j]][0] - centroids[i][0]) + (X2d[row_labels[i][j]][1] - centroids[i][1])*(X2d[row_labels[i][j]][1] - centroids[i][1]) < min_dist:\n",
    "                min_dist = (X2d[row_labels[i][j]][0] - centroids[i][0])*(X2d[row_labels[i][j]][0] - centroids[i][0]) + (X2d[row_labels[i][j]][1] - centroids[i][1])*(X2d[row_labels[i][j]][1] - centroids[i][1])\n",
    "                ind_min = row_labels[i][j]\n",
    "        average_clients.append(ind_min)\n",
    "    return average_clients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 426,
   "metadata": {},
   "outputs": [],
   "source": [
    "X2d_france_arr = X2d_france.to_numpy()\n",
    "X2d_japan_arr = X2d_japan.to_numpy()\n",
    "X2d_us_arr = X2d_us.to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 427,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Getting number of clusters that maximizes silhouette score\n",
    "c_france = maximize_cluster_score(X2d_france, 3, 30)\n",
    "c_japan = maximize_cluster_score(X2d_japan, 3, 30)\n",
    "c_us = maximize_cluster_score(X2d_us, 3, 30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 428,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9, 0.9981490564462092) (8, 0.9998422306115123) (11, 0.9975237156053496)\n"
     ]
    }
   ],
   "source": [
    "print(c_france, c_japan, c_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 429,
   "metadata": {},
   "outputs": [],
   "source": [
    "average_clients_france = get_average_clients(df_FRANCE_Apparel, X2d_france_arr, c1)\n",
    "average_clients_japan = get_average_clients(df_JAPAN_Apparel, X2d_japan_arr, c2)\n",
    "average_clients_us = get_average_clients(df_UNITED_STATES_Apparel, X2d_us_arr, c3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 387,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Getting indices for clients in the original dataset\n",
    "def get_index_client(dict_var, expected_var):\n",
    "    index_client = []\n",
    "    for i in range(len(expected_var)):\n",
    "        index_client.append(dict_var[expected_var[i]])\n",
    "    return index_client"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 388,
   "metadata": {},
   "outputs": [],
   "source": [
    "ind_exp_france = get_index_client(dict_france, average_clients_france)\n",
    "ind_exp_japan = get_index_client(dict_japan, average_clients_japan)\n",
    "ind_exp_us = get_index_client(dict_us, average_clients_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[7923, 9524, 10303, 1540]"
      ]
     },
     "execution_count": 389,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind_exp_france"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 390,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[15937, 14677, 12847, 1411]"
      ]
     },
     "execution_count": 390,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind_exp_japan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 391,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[8721, 14139, 14398, 12039, 4142]"
      ]
     },
     "execution_count": 391,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind_exp_us"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 392,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Get clients in the original dataset\n",
    "def get_clients(ind_exp):\n",
    "    clients = []\n",
    "    for i in range(len(ind_exp)):\n",
    "        clients.append(X['unique_client'][ind_exp[i]])\n",
    "    return clients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 393,
   "metadata": {},
   "outputs": [],
   "source": [
    "clients_france = get_clients(ind_exp_france)\n",
    "clients_japan = get_clients(ind_exp_japan)\n",
    "clients_us = get_clients(ind_exp_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 394,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ga_a2c851c7-7c38-e211-9847-000c29ae377c_FRANCE',\n",
       " 'ga_1acf51c7-7c38-e211-9847-000c29ae377c_FRANCE',\n",
       " 'ga_db07009b-ed73-e511-9606-005056bb2344_FRANCE',\n",
       " '962e4ce2-4b06-3c00-eba7-1a2717783027_FRANCE']"
      ]
     },
     "execution_count": 394,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clients_france"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 395,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ga_8fd2bab0-9fb5-e511-9606-005056bb2344_JAPAN',\n",
       " 'ga_da1733da-7c38-e211-9847-000c29ae377c_JAPAN',\n",
       " 'c9dcf697-5b42-bea2-4410-08634a484fed_JAPAN',\n",
       " 'ga_98ad51c7-7c38-e211-9847-000c29ae377c_JAPAN']"
      ]
     },
     "execution_count": 395,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clients_japan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 396,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ga_a5cce4c2-33a6-e311-b5d4-005056bb7e02_UNITED STATES',\n",
       " 'ga_51106280-082b-e511-9606-005056bb2344_UNITED STATES',\n",
       " 'ga_8e7a133d-fbda-e511-9606-005056bb2344_UNITED STATES',\n",
       " 'ga_506bb33e-88a5-e411-b0d2-005056bb7e02_UNITED STATES',\n",
       " 'eb63f8dc-5363-c395-eb56-ea2246c5cf21_UNITED STATES']"
      ]
     },
     "execution_count": 396,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clients_us"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 397,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_france_clients = X.iloc[ind_exp_france,:]\n",
    "df_japan_clients = X.iloc[ind_exp_japan,:]\n",
    "df_us_clients = X.iloc[ind_exp_us,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_france_clients.to_csv('france_clients.csv')\n",
    "df_japan_clients.to_csv('japan_clients.csv')\n",
    "df_us_clients.to_csv('us_clients.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get best client in each cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 399,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_FRANCE_Apparel = df_FRANCE_Apparel.fillna(0)\n",
    "for i in range(df_FRANCE_Apparel.shape[0]):\n",
    "    df_FRANCE_Apparel.loc[i,'duration'] = float(df_FRANCE_Apparel['churn_date_pred'][i]) - float(df_FRANCE_Apparel['time'][i])\n",
    "\n",
    "    \n",
    "df_JAPAN_Apparel = df_JAPAN_Apparel.fillna(0)\n",
    "for i in range(df_JAPAN_Apparel.shape[0]):\n",
    "    df_JAPAN_Apparel.loc[i,'duration'] = float(df_JAPAN_Apparel['churn_date_pred'][i]) - float(df_JAPAN_Apparel['time'][i])\n",
    "    \n",
    "    \n",
    "df_UNITED_STATES_Apparel = df_UNITED_STATES_Apparel.fillna(0)\n",
    "for i in range(df_UNITED_STATES_Apparel.shape[0]):\n",
    "    df_UNITED_STATES_Apparel.loc[i,'duration'] = float(df_UNITED_STATES_Apparel['churn_date_pred'][i]) - float(df_UNITED_STATES_Apparel['time'][i])\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 400,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rep_clients_per_cluster_FR(df_FRANCE_Apparel):\n",
    "    best_clients = []\n",
    "    labels = list(df_FRANCE_Apparel['labels'])\n",
    "    set_labels = set(labels)\n",
    "    labels_quant = len(set_labels)\n",
    "    for i in range(labels_quant):\n",
    "        df_cluster = df_FRANCE_Apparel[df_FRANCE_Apparel['labels'] == i]\n",
    "        min_churn_date = -1\n",
    "        max_churn_date = -1\n",
    "        ind_min = 0\n",
    "        ind_max = 0\n",
    "        for j, row in df_cluster.iterrows():\n",
    "            if df_cluster['duration'][j] > max_churn_date:\n",
    "                max_churn_date = df_cluster['duration'][j]\n",
    "                ind_max = j\n",
    "            if min_churn_date == -1 or df_cluster['duration'][j] < min_churn_date:\n",
    "                min_churn_date = df_cluster['duration'][j]\n",
    "                ind_min = j\n",
    "        best_clients.append(ind_max)\n",
    "    return (best_clients)\n",
    "\n",
    "\n",
    "def rep_clients_per_cluster_JP(df_JAPAN_Apparel):\n",
    "    best_clients = []\n",
    "    labels = list(df_JAPAN_Apparel['labels'])\n",
    "    set_labels = set(labels)\n",
    "    labels_quant = len(set_labels)\n",
    "    for i in range(labels_quant):\n",
    "        df_cluster = df_JAPAN_Apparel[df_JAPAN_Apparel['labels'] == i]\n",
    "        min_churn_date = -1\n",
    "        max_churn_date = -1\n",
    "        ind_min = 0\n",
    "        ind_max = 0\n",
    "        for j, row in df_cluster.iterrows():\n",
    "            if df_cluster['duration'][j] > max_churn_date:\n",
    "                max_churn_date = df_cluster['duration'][j]\n",
    "                ind_max = j\n",
    "            if min_churn_date == -1 or df_cluster['duration'][j] < min_churn_date:\n",
    "                min_churn_date = df_cluster['duration'][j]\n",
    "                ind_min = j\n",
    "        best_clients.append(ind_max)\n",
    "    return (best_clients)\n",
    "\n",
    "def rep_clients_per_cluster_US(df_UNITED_STATES_Apparel):\n",
    "    best_clients = []\n",
    "    labels = list(df_UNITED_STATES_Apparel['labels'])\n",
    "    set_labels = set(labels)\n",
    "    labels_quant = len(set_labels)\n",
    "    for i in range(labels_quant):\n",
    "        df_cluster = df_UNITED_STATES_Apparel[df_UNITED_STATES_Apparel['labels'] == i]\n",
    "        min_churn_date = -1\n",
    "        max_churn_date = -1\n",
    "        ind_min = 0\n",
    "        ind_max = 0\n",
    "        for j, row in df_cluster.iterrows():\n",
    "            if df_cluster['duration'][j] > max_churn_date:\n",
    "                max_churn_date = df_cluster['duration'][j]\n",
    "                ind_max = j\n",
    "            if min_churn_date == -1 or df_cluster['duration'][j] < min_churn_date:\n",
    "                min_churn_date = df_cluster['duration'][j]\n",
    "                ind_min = j\n",
    "        best_clients.append(ind_max)\n",
    "    return (best_clients)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 401,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_clients_france = rep_clients_per_cluster_FR(df_FRANCE_Apparel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 402,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[146, 37, 539, 44]"
      ]
     },
     "execution_count": 402,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Indices in the cluster dataframe\n",
    "best_clients_france"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_clients_japan = rep_clients_per_cluster_JP(df_JAPAN_Apparel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 404,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[260, 90, 399, 140]"
      ]
     },
     "execution_count": 404,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Indices in the cluster dataframe\n",
    "best_clients_japan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 405,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_clients_us = rep_clients_per_cluster_US(df_UNITED_STATES_Apparel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 406,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[731, 367, 1208, 28, 320]"
      ]
     },
     "execution_count": 406,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Indices in the cluster dataframe\n",
    "best_clients_us"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 407,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Indices in the original dataframe\n",
    "ind_best_france = get_index_client(dict_france, best_clients_france)\n",
    "ind_best_japan = get_index_client(dict_japan, best_clients_japan)\n",
    "ind_best_us = get_index_client(dict_us, best_clients_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 408,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[4632, 1326, 16023, 1540]"
      ]
     },
     "execution_count": 408,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind_best_france"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Append clients IDs to he original dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 409,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['5a9f9ccf-e847-9940-4823-ce747bd2e6d2_FRANCE',\n",
       " 'ga_cc58808f-1ee3-e411-9606-005056bb2344_FRANCE',\n",
       " 'ga_3d322f95-123b-e211-9847-000c29ae377c_FRANCE',\n",
       " '962e4ce2-4b06-3c00-eba7-1a2717783027_FRANCE']"
      ]
     },
     "execution_count": 409,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_clients(ind_best_france)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 410,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['8df3cd43-e950-7919-30f7-169a5758563a_JAPAN',\n",
       " 'ga_7cc451c7-7c38-e211-9847-000c29ae377c_JAPAN',\n",
       " 'ga_3d322f95-123b-e211-9847-000c29ae377c_JAPAN',\n",
       " '1f04248e-5506-30d7-dabf-27eee12f8ad2_JAPAN']"
      ]
     },
     "execution_count": 410,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_clients(ind_best_japan)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 411,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ga_e2824d59-5d7a-e411-b0d2-005056bb7e02_UNITED STATES',\n",
       " 'ga_96bc51c7-7c38-e211-9847-000c29ae377c_UNITED STATES',\n",
       " 'ga_3d322f95-123b-e211-9847-000c29ae377c_UNITED STATES',\n",
       " 'ga_ed7d17e4-aa86-e211-a813-005056bb2836_UNITED STATES',\n",
       " 'eb63f8dc-5363-c395-eb56-ea2246c5cf21_UNITED STATES']"
      ]
     },
     "execution_count": 411,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_clients(ind_best_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 412,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_france_best = X.iloc[ind_best_france,:]\n",
    "df_japan_best = X.iloc[ind_best_japan,:]\n",
    "df_us_best = X.iloc[ind_best_us,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 413,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_france_best.to_csv('france_best.csv')\n",
    "df_japan_best.to_csv('japan_best.csv')\n",
    "df_us_best.to_csv('us_best.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Indicate the difference between the best and the average client by country, indusrty, cluster, to suggest next strategy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 459,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['unique_client', 'month', 'time', 'vertical', 'industry',\n",
       "       'client_country', 'client_continent', 'client_region', 'OS', 'device',\n",
       "       'environment', 'criteo_product', 'spend_criteo_euro', 'sales_criteo',\n",
       "       'CPO', 'revenue_criteo_euro', 'AOV', 'COS', 'Android', 'MacOS', 'Other',\n",
       "       'iOs', 'Desktop', 'Mobile', 'app', 'web', 'awareness', 'consideration',\n",
       "       'conversion', 'products', 'onsite_sales', 'onsite_revenue_euro',\n",
       "       'churn_date', 'country_industry'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 459,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 460,
   "metadata": {},
   "outputs": [],
   "source": [
    "df3 = X[['spend_criteo_euro', 'sales_criteo',\n",
    "       'CPO', 'revenue_criteo_euro', 'AOV', 'COS', 'Android', 'MacOS', 'Other',\n",
    "       'iOs', 'Desktop', 'Mobile', 'app', 'web', 'awareness', 'consideration',\n",
    "       'conversion', 'products', 'onsite_sales', 'onsite_revenue_euro',\n",
    "       'country_industry']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 465,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['spend_criteo_euro', 'sales_criteo', 'CPO', 'revenue_criteo_euro',\n",
       "       'AOV', 'COS', 'Android', 'MacOS', 'Other', 'iOs', 'Desktop', 'Mobile',\n",
       "       'app', 'web', 'awareness', 'consideration', 'conversion', 'products',\n",
       "       'onsite_sales', 'onsite_revenue_euro', 'country_industry'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 465,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df3.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 466,
   "metadata": {},
   "outputs": [],
   "source": [
    "def indicate_changements(df3, average_clients, best_clients):\n",
    "    for i in range(len(average_clients)):\n",
    "        print('Difference from the Best Client ', i)\n",
    "        for column in df3:\n",
    "            if column != 'client_country' and column != 'client_region' and column != 'time' and column != 'unique_client' and column != 'churn_date':\n",
    "                if df3[column][average_clients[i]] != df3[column][best_clients[i]]:\n",
    "                    print('change value', column, 'to ', df3[column][best_clients[i]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 467,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference from the Best Client  0\n",
      "change value index to  4632\n",
      "change value spend_criteo_euro to  107523\n",
      "change value sales_criteo to  4485\n",
      "change value CPO to  24.0\n",
      "change value revenue_criteo_euro to  420841\n",
      "change value AOV to  94.0\n",
      "change value Android to  107523.0\n",
      "change value Mobile to  107523.0\n",
      "change value web to  107523.0\n",
      "change value conversion to  107523.0\n",
      "change value products to  3228.0\n",
      "change value onsite_sales to  10019.0\n",
      "change value onsite_revenue_euro to  27906488.35\n",
      "change value churn_date_pred to  2024.07\n",
      "change value duration to  7.039999999999964\n",
      "Difference from the Best Client  1\n",
      "change value index to  1326\n",
      "change value month to  12\n",
      "change value environment to  app\n",
      "change value criteo_product to  conversion\n",
      "change value spend_criteo_euro to  20\n",
      "change value sales_criteo to  12\n",
      "change value CPO to  2.0\n",
      "change value revenue_criteo_euro to  757\n",
      "change value AOV to  63.0\n",
      "change value Android to  20.0\n",
      "change value Mobile to  20.0\n",
      "change value app to  20.0\n",
      "change value web to  0.0\n",
      "change value consideration to  0.0\n",
      "change value conversion to  20.0\n",
      "change value products to  19019.0\n",
      "change value onsite_sales to  4.0\n",
      "change value onsite_revenue_euro to  127.96\n",
      "change value churn_date_pred to  2024.12\n",
      "change value duration to  6.0\n",
      "Difference from the Best Client  2\n",
      "change value index to  16023\n",
      "change value OS to  Android\n",
      "change value device to  Mobile\n",
      "change value environment to  app\n",
      "change value criteo_product to  consideration\n",
      "change value spend_criteo_euro to  12\n",
      "change value sales_criteo to  125\n",
      "change value CPO to  0.0\n",
      "change value revenue_criteo_euro to  13907\n",
      "change value AOV to  111.0\n",
      "change value Android to  12.0\n",
      "change value MacOS to  0.0\n",
      "change value Desktop to  0.0\n",
      "change value Mobile to  12.0\n",
      "change value app to  12.0\n",
      "change value web to  0.0\n",
      "change value consideration to  12.0\n",
      "change value conversion to  0.0\n",
      "change value products to  1808794.0\n",
      "change value onsite_sales to  3167.0\n",
      "change value onsite_revenue_euro to  44997.21\n",
      "change value churn_date_pred to  2024.05\n",
      "change value duration to  7.019999999999982\n",
      "Difference from the Best Client  3\n"
     ]
    }
   ],
   "source": [
    "indicate_changements(df_FRANCE_Apparel, average_clients_france, best_clients_france)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 463,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference from the Best Client  0\n",
      "change calue index to  12030\n",
      "change calue month to  8\n",
      "change calue spend_criteo_euro to  21196\n",
      "change calue sales_criteo to  1424\n",
      "change calue CPO to  15.0\n",
      "change calue revenue_criteo_euro to  8\n",
      "change calue AOV to  0.0\n",
      "change calue COS to  2717\n",
      "change calue Android to  21196.0\n",
      "change calue Mobile to  21196.0\n",
      "change calue web to  21196.0\n",
      "change calue conversion to  21196.0\n",
      "change calue products to  31859.0\n",
      "change calue onsite_sales to  136.0\n",
      "change calue onsite_revenue_euro to  121635.1\n",
      "change calue churn_date_pred to  2025.06\n",
      "change calue duration to  5.980000000000018\n",
      "Difference from the Best Client  1\n",
      "change calue index to  3749\n",
      "change calue spend_criteo_euro to  106166\n",
      "change calue sales_criteo to  7742\n",
      "change calue CPO to  14.0\n",
      "change calue revenue_criteo_euro to  481812\n",
      "change calue AOV to  62.0\n",
      "change calue Android to  106166.0\n",
      "change calue Mobile to  106166.0\n",
      "change calue web to  106166.0\n",
      "change calue conversion to  106166.0\n",
      "change calue products to  7915.0\n",
      "change calue onsite_sales to  827.0\n",
      "change calue onsite_revenue_euro to  3805.56\n",
      "change calue churn_date_pred to  2025.02\n",
      "change calue duration to  7.990000000000009\n",
      "Difference from the Best Client  2\n",
      "change calue index to  16026\n",
      "change calue month to  9\n",
      "change calue OS to  Android\n",
      "change calue device to  Mobile\n",
      "change calue environment to  app\n",
      "change calue criteo_product to  consideration\n",
      "change calue spend_criteo_euro to  8\n",
      "change calue sales_criteo to  4\n",
      "change calue CPO to  2.0\n",
      "change calue revenue_criteo_euro to  332\n",
      "change calue AOV to  83.0\n",
      "change calue COS to  0\n",
      "change calue Android to  8.0\n",
      "change calue MacOS to  0.0\n",
      "change calue Desktop to  0.0\n",
      "change calue Mobile to  8.0\n",
      "change calue app to  8.0\n",
      "change calue web to  0.0\n",
      "change calue consideration to  8.0\n",
      "change calue conversion to  0.0\n",
      "change calue products to  1372339.0\n",
      "change calue onsite_sales to  74.0\n",
      "change calue onsite_revenue_euro to  24152.54\n",
      "change calue churn_date_pred to  2024.07\n",
      "change calue duration to  6.980000000000018\n",
      "Difference from the Best Client  3\n",
      "change calue index to  6217\n",
      "change calue month to  12\n",
      "change calue spend_criteo_euro to  6396\n",
      "change calue sales_criteo to  4\n",
      "change calue CPO to  1599.0\n",
      "change calue revenue_criteo_euro to  581\n",
      "change calue AOV to  145.0\n",
      "change calue COS to  11\n",
      "change calue Android to  6396.0\n",
      "change calue Mobile to  6396.0\n",
      "change calue web to  6396.0\n",
      "change calue consideration to  6396.0\n",
      "change calue products to  3427.0\n",
      "change calue onsite_sales to  111263.0\n",
      "change calue onsite_revenue_euro to  7701518.21\n",
      "change calue churn_date_pred to  2026.03\n",
      "change calue duration to  6.910000000000082\n"
     ]
    }
   ],
   "source": [
    "indicate_changements(df_JAPAN_Apparel, average_clients_japan, best_clients_japan)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 464,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference from the Best Client  0\n",
      "change calue index to  10121\n",
      "change calue spend_criteo_euro to  149920\n",
      "change calue sales_criteo to  238\n",
      "change calue CPO to  630.0\n",
      "change calue revenue_criteo_euro to  237851\n",
      "change calue AOV to  999.0\n",
      "change calue Android to  149920.0\n",
      "change calue Mobile to  149920.0\n",
      "change calue web to  149920.0\n",
      "change calue conversion to  149920.0\n",
      "change calue products to  59918.0\n",
      "change calue onsite_sales to  199.0\n",
      "change calue onsite_revenue_euro to  14972.98\n",
      "change calue churn_date_pred to  2025.02\n",
      "change calue duration to  7.990000000000009\n",
      "Difference from the Best Client  1\n",
      "change calue index to  4913\n",
      "change calue month to  5\n",
      "change calue device to  Mobile\n",
      "change calue environment to  app\n",
      "change calue spend_criteo_euro to  3409\n",
      "change calue sales_criteo to  62\n",
      "change calue CPO to  55.0\n",
      "change calue revenue_criteo_euro to  30265\n",
      "change calue AOV to  488.0\n",
      "change calue Android to  3409.0\n",
      "change calue Desktop to  0.0\n",
      "change calue Mobile to  3409.0\n",
      "change calue app to  3409.0\n",
      "change calue web to  0.0\n",
      "change calue conversion to  3409.0\n",
      "change calue products to  3054.0\n",
      "change calue onsite_sales to  675.0\n",
      "change calue onsite_revenue_euro to  1462460.7\n",
      "change calue churn_date_pred to  2025.04\n",
      "change calue duration to  6.990000000000009\n",
      "Difference from the Best Client  2\n",
      "change calue index to  16038\n",
      "change calue month to  11\n",
      "change calue OS to  Android\n",
      "change calue device to  Mobile\n",
      "change calue environment to  app\n",
      "change calue criteo_product to  consideration\n",
      "change calue spend_criteo_euro to  2079\n",
      "change calue sales_criteo to  273\n",
      "change calue CPO to  8.0\n",
      "change calue revenue_criteo_euro to  18498\n",
      "change calue AOV to  68.0\n",
      "change calue COS to  0\n",
      "change calue Android to  2079.0\n",
      "change calue MacOS to  0.0\n",
      "change calue Desktop to  0.0\n",
      "change calue Mobile to  2079.0\n",
      "change calue app to  2079.0\n",
      "change calue web to  0.0\n",
      "change calue consideration to  2079.0\n",
      "change calue conversion to  0.0\n",
      "change calue products to  1511883.0\n",
      "change calue onsite_sales to  94.0\n",
      "change calue onsite_revenue_euro to  34115.29\n",
      "change calue churn_date_pred to  2024.1\n",
      "change calue duration to  6.990000000000009\n",
      "Difference from the Best Client  3\n",
      "change calue index to  406\n",
      "change calue month to  8\n",
      "change calue spend_criteo_euro to  43\n",
      "change calue CPO to  11.0\n",
      "change calue revenue_criteo_euro to  441\n",
      "change calue AOV to  110.0\n",
      "change calue Android to  43.0\n",
      "change calue Mobile to  43.0\n",
      "change calue web to  43.0\n",
      "change calue conversion to  43.0\n",
      "change calue products to  3478.0\n",
      "change calue onsite_sales to  325100.0\n",
      "change calue onsite_revenue_euro to  26425593.09\n",
      "change calue churn_date_pred to  2022.09\n",
      "change calue duration to  4.009999999999991\n",
      "Difference from the Best Client  4\n"
     ]
    }
   ],
   "source": [
    "indicate_changements(df_UNITED_STATES_Apparel, average_clients_us, best_clients_us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}