analytics_and_recommendation_engine.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.container { width:100% !important; }</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import datetime\n",
    "\n",
    "import scipy as sp\n",
    "from scipy import stats\n",
    "\n",
    "from IPython.core.display import display, HTML\n",
    "display(HTML(\"<style>.container { width:100% !important; }</style>\"))\n",
    "\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format='retina'\n",
    "\n",
    "pd.set_option('display.max_rows', None)\n",
    "pd.set_option('display.max_columns', None)\n",
    "pd.set_option('display.width', None)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "from IPython.display import display_html\n",
    "from itertools import chain,cycle\n",
    "def display_side_by_side(*args,titles=cycle([''])):\n",
    "    html_str=''\n",
    "    for df,title in zip(args, chain(titles,cycle(['</br>'])) ):\n",
    "        html_str+='<th style=\"text-align:center\"><td style=\"vertical-align:top\">'\n",
    "        html_str+=f'<h2>{title}</h2>'\n",
    "        html_str+=df.to_html(index=False).replace('table','table style=\"display:inline\"')\n",
    "        html_str+='</td></th>'\n",
    "    display_html(html_str,raw=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Product recommendations.png\r\n",
      "README.md\r\n",
      "analytics_and_recommendation_engine.ipynb\r\n",
      "e-commerce_data.csv\r\n"
     ]
    }
   ],
   "source": [
    "!ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using RFM Segmentation (Recency, Frequency and Monetary Value Analysis) to recommend other purchases:\n",
    "\n",
    "### First step:\n",
    "    \n",
    "**RFM segmentation is a great method to identify groups of customers for special treatment. Excellent method to improve your customer marketing.**\n",
    "\n",
    "While there are countless ways to perform segmentation, RFM analysis is popular for three reasons:\n",
    "\n",
    "It utilizes objective, numerical scales that yield a concise and informative high-level depiction of customers.\n",
    "It is simple – marketers can use it effectively without the need for data scientists or sophisticated software.\n",
    "It is intuitive – the output of this segmentation method is easy to understand and interpret.\n",
    "\n",
    "- **Recency:** How much time has elapsed since a customer’s last activity or transaction with the brand? (Activity could be the last visit to a website or use of a mobile app, etc...)\n",
    "\n",
    "- **Frequency:** How often has a customer transacted or interacted with the brand during a particular period of time? Clearly, customers with frequent activities are more engaged, and probably more loyal, than customers who rarely do so. And one-time-only customers are in a class of their own.\n",
    "\n",
    "- **Monetary:** Also referred to as “monetary value,” this factor reflects how much a customer has spent with the brand during a particular period of time. Big spenders should usually be treated differently than customers who spend little. Looking at monetary divided by frequency indicates the average purchase amount – an important secondary factor to consider when segmenting customers.\n",
    "\n",
    "\n",
    "### Second step:\n",
    "\n",
    "Using Singular Value Decomposition (SVD) to recommend purchases using our best customers identified in the above step for better targeting.\n",
    "\n",
    "SVD is a classical Collaborative Filtering method from linear algebra popular for developing recommender systems.\n",
    "\n",
    "#### What Is Collaborative Filtering?\n",
    "\n",
    "Collaborative filtering is a technique that can filter out items that a user might like on the basis of reactions by similar users.\n",
    "\n",
    "It works by searching a large group of people and finding a smaller set of users with tastes similar to a particular user. It looks at the items they like and combines them to create a ranked list of suggestions.\n",
    "\n",
    "\n",
    "### Third step:\n",
    "\n",
    "Using Dimensionality reduction using truncated Singular Value Decomposition (truncatedSVD) with a correlation matrix to recommend other products that customers also bought with a specific product. AKA The \"people who bought this also bought...\" that you see in your cart before finalising your purchase.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load and prepare dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 541909 columns: 8\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>InvoiceNo</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>UnitPrice</th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>Country</th>\n",
       "      <th>TotalPrice</th>\n",
       "      <th>day</th>\n",
       "      <th>day_name</th>\n",
       "      <th>day_of_year</th>\n",
       "      <th>week</th>\n",
       "      <th>hour</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>weekday</th>\n",
       "      <th>weekend</th>\n",
       "      <th>month_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>536365</td>\n",
       "      <td>85123A</td>\n",
       "      <td>WHITE HANGING HEART T-LIGHT HOLDER</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>2.55</td>\n",
       "      <td>17850</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>15.30</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>536365</td>\n",
       "      <td>71053</td>\n",
       "      <td>WHITE METAL LANTERN</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>3.39</td>\n",
       "      <td>17850</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>20.34</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>536365</td>\n",
       "      <td>84406B</td>\n",
       "      <td>CREAM CUPID HEARTS COAT HANGER</td>\n",
       "      <td>8</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>2.75</td>\n",
       "      <td>17850</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>22.00</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  InvoiceNo StockCode                         Description  Quantity  \\\n",
       "0    536365    85123A  WHITE HANGING HEART T-LIGHT HOLDER         6   \n",
       "1    536365     71053                 WHITE METAL LANTERN         6   \n",
       "2    536365    84406B      CREAM CUPID HEARTS COAT HANGER         8   \n",
       "\n",
       "          InvoiceDate  UnitPrice CustomerID         Country  TotalPrice  day  \\\n",
       "0 2010-12-01 08:26:00       2.55      17850  United Kingdom       15.30    1   \n",
       "1 2010-12-01 08:26:00       3.39      17850  United Kingdom       20.34    1   \n",
       "2 2010-12-01 08:26:00       2.75      17850  United Kingdom       22.00    1   \n",
       "\n",
       "    day_name  day_of_year  week  hour  month  year  weekday  weekend  \\\n",
       "0  Wednesday          335    48     8     12  2010        2        0   \n",
       "1  Wednesday          335    48     8     12  2010        2        0   \n",
       "2  Wednesday          335    48     8     12  2010        2        0   \n",
       "\n",
       "  month_year  \n",
       "0    2010-12  \n",
       "1    2010-12  \n",
       "2    2010-12  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('e-commerce_data.csv',encoding = \"ISO-8859-1\")\n",
    "print('rows:',df.shape[0],'columns:',df.shape[1])\n",
    "\n",
    "df['TotalPrice'] = df['Quantity']*df['UnitPrice']\n",
    "df['CustomerID'] = df['CustomerID'].fillna(0).astype(int).astype(str).replace('0',np.nan)\n",
    "df['InvoiceDate'] = pd.to_datetime(df['InvoiceDate'])\n",
    "\n",
    "df['day'] = df['InvoiceDate'].dt.day\n",
    "df['day_name'] = df['InvoiceDate'].dt.day_name() \n",
    "df['day_of_year'] = df['InvoiceDate'].dt.dayofyear \n",
    "df['week'] = df['InvoiceDate'].dt.weekofyear \n",
    "df['hour'] = df['InvoiceDate'].dt.hour  \n",
    "df['month'] = df['InvoiceDate'].dt.month\n",
    "df['year'] = df['InvoiceDate'].dt.year\n",
    "df['weekday'] = df['InvoiceDate'].dt.dayofweek\n",
    "df['weekend'] = [0 if x <= 5 else 1 for x in df['weekday']]\n",
    "\n",
    "df['month_year'] = df['year'].map(str)+ '-' +df['month'].map(str)\n",
    "df['month_year'] = pd.to_datetime(df['month_year'], format='%Y-%m').dt.strftime('%Y-%m')\n",
    "\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset Shape: (541909, 19)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>dtypes</th>\n",
       "      <th>Missing</th>\n",
       "      <th>Missing_%</th>\n",
       "      <th>Uniques</th>\n",
       "      <th>First 5 Unique Values</th>\n",
       "      <th>Entropy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>InvoiceNo</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>25900</td>\n",
       "      <td>[536365, 536366, 536367, 536368, 536369]</td>\n",
       "      <td>13.39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>StockCode</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>4070</td>\n",
       "      <td>[85123A, 71053, 84406B, 84029G, 84029E]</td>\n",
       "      <td>10.91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Description</td>\n",
       "      <td>object</td>\n",
       "      <td>1454</td>\n",
       "      <td>0.27</td>\n",
       "      <td>4223</td>\n",
       "      <td>[WHITE HANGING HEART T-LIGHT HOLDER, WHITE MET...</td>\n",
       "      <td>10.93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Quantity</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>722</td>\n",
       "      <td>[6, 8, 2, 32, 3]</td>\n",
       "      <td>3.72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>InvoiceDate</td>\n",
       "      <td>datetime64[ns]</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>23260</td>\n",
       "      <td>[2010-12-01T08:26:00.000000000, 2010-12-01T08:...</td>\n",
       "      <td>13.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>UnitPrice</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1630</td>\n",
       "      <td>[2.55, 3.39, 2.75, 7.65, 4.25]</td>\n",
       "      <td>5.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>CustomerID</td>\n",
       "      <td>object</td>\n",
       "      <td>135080</td>\n",
       "      <td>24.93</td>\n",
       "      <td>4372</td>\n",
       "      <td>[17850, 13047, 12583, 13748, 15100]</td>\n",
       "      <td>10.85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Country</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>38</td>\n",
       "      <td>[United Kingdom, France, Australia, Netherland...</td>\n",
       "      <td>0.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>TotalPrice</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6204</td>\n",
       "      <td>[15.299999999999999, 20.34, 22.0, 15.3, 25.5]</td>\n",
       "      <td>8.38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>day</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>31</td>\n",
       "      <td>[1, 2, 3, 5, 6]</td>\n",
       "      <td>4.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>day_name</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6</td>\n",
       "      <td>[Wednesday, Thursday, Friday, Sunday, Monday]</td>\n",
       "      <td>2.57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>day_of_year</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>298</td>\n",
       "      <td>[335, 336, 337, 339, 340]</td>\n",
       "      <td>8.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>week</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>51</td>\n",
       "      <td>[48, 49, 50, 51, 1]</td>\n",
       "      <td>5.52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>hour</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>15</td>\n",
       "      <td>[8, 9, 10, 11, 12]</td>\n",
       "      <td>3.29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>month</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>12</td>\n",
       "      <td>[12, 1, 2, 3, 4]</td>\n",
       "      <td>3.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>year</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>[2010, 2011]</td>\n",
       "      <td>0.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>weekday</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6</td>\n",
       "      <td>[2, 3, 4, 6, 0]</td>\n",
       "      <td>2.57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>weekend</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>[0, 1]</td>\n",
       "      <td>0.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>month_year</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>13</td>\n",
       "      <td>[2010-12, 2011-01, 2011-02, 2011-03, 2011-04]</td>\n",
       "      <td>3.62</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Name          dtypes  Missing  Missing_%  Uniques  \\\n",
       "0     InvoiceNo          object        0       0.00    25900   \n",
       "1     StockCode          object        0       0.00     4070   \n",
       "2   Description          object     1454       0.27     4223   \n",
       "3      Quantity           int64        0       0.00      722   \n",
       "4   InvoiceDate  datetime64[ns]        0       0.00    23260   \n",
       "5     UnitPrice         float64        0       0.00     1630   \n",
       "6    CustomerID          object   135080      24.93     4372   \n",
       "7       Country          object        0       0.00       38   \n",
       "8    TotalPrice         float64        0       0.00     6204   \n",
       "9           day           int64        0       0.00       31   \n",
       "10     day_name          object        0       0.00        6   \n",
       "11  day_of_year           int64        0       0.00      298   \n",
       "12         week           int64        0       0.00       51   \n",
       "13         hour           int64        0       0.00       15   \n",
       "14        month           int64        0       0.00       12   \n",
       "15         year           int64        0       0.00        2   \n",
       "16      weekday           int64        0       0.00        6   \n",
       "17      weekend           int64        0       0.00        2   \n",
       "18   month_year          object        0       0.00       13   \n",
       "\n",
       "                                First 5 Unique Values  Entropy  \n",
       "0            [536365, 536366, 536367, 536368, 536369]    13.39  \n",
       "1             [85123A, 71053, 84406B, 84029G, 84029E]    10.91  \n",
       "2   [WHITE HANGING HEART T-LIGHT HOLDER, WHITE MET...    10.93  \n",
       "3                                    [6, 8, 2, 32, 3]     3.72  \n",
       "4   [2010-12-01T08:26:00.000000000, 2010-12-01T08:...    13.28  \n",
       "5                      [2.55, 3.39, 2.75, 7.65, 4.25]     5.68  \n",
       "6                 [17850, 13047, 12583, 13748, 15100]    10.85  \n",
       "7   [United Kingdom, France, Australia, Netherland...     0.74  \n",
       "8       [15.299999999999999, 20.34, 22.0, 15.3, 25.5]     8.38  \n",
       "9                                     [1, 2, 3, 5, 6]     4.92  \n",
       "10      [Wednesday, Thursday, Friday, Sunday, Monday]     2.57  \n",
       "11                          [335, 336, 337, 339, 340]     8.03  \n",
       "12                                [48, 49, 50, 51, 1]     5.52  \n",
       "13                                 [8, 9, 10, 11, 12]     3.29  \n",
       "14                                   [12, 1, 2, 3, 4]     3.50  \n",
       "15                                       [2010, 2011]     0.40  \n",
       "16                                    [2, 3, 4, 6, 0]     2.57  \n",
       "17                                             [0, 1]     0.53  \n",
       "18      [2010-12, 2011-01, 2011-02, 2011-03, 2011-04]     3.62  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def resume_table(df):\n",
    "    print(f\"Dataset Shape: {df.shape}\")\n",
    "    summary = pd.DataFrame(df.dtypes,columns=['dtypes'])\n",
    "    summary = summary.reset_index()\n",
    "    summary['Name'] = summary['index']\n",
    "    summary = summary[['Name','dtypes']]\n",
    "    summary['Missing'] = df.isnull().sum().values \n",
    "    summary['Missing_%'] = np.round(df.isnull().sum().values / len(df) *100,2)\n",
    "    summary['Uniques'] = df.nunique().values\n",
    "    summary['First 5 Unique Values'] = [x for x in pd.Series({c: df[c].unique()[:5] for c in df})]\n",
    "    \n",
    "    ## Entropy is the measures of impurity, disorder or uncertainty.\n",
    "    ## Entropy controls how a Decision Tree decides to split the data. It actually effects how a Decision Tree draws its boundaries.\n",
    "    ## An attribute with highest Information gain will tested/split first.\n",
    "    for name in summary['Name'].value_counts().index:\n",
    "        summary.loc[summary['Name'] == name, 'Entropy'] = round(stats.entropy(df[name].value_counts(normalize=True), base=2),2) \n",
    "\n",
    "    return summary\n",
    "\n",
    "df_resume = resume_table(df)\n",
    "df_resume"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df[df['InvoiceNo'].str.contains('C')]['InvoiceNo'].unique()\n",
    "df = df[~df['InvoiceNo'].str.contains('C')]\n",
    "df = df[df['Description'] != 'POSTAGE']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Columns 'Description' and 'CustomerID' have missing values:\n",
    "\n",
    "- 'Description' feature not only have empty descriptions but also incorrect ones (same stockcode with different descriptions), so instead of just dropping them we will map them back based on their relation 'StockCode' to 'Description'.\n",
    "\n",
    "- 'CustomerID' feature has 3710 invoice with no CustomerID which represents 135 thousand lines, I will filter them out for this exercise."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Treatment of the 'Description' column:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Empty descriptions before fix: 1454\n",
      "Empty descriptions after fix: 0\n"
     ]
    }
   ],
   "source": [
    "df['Description'] = df['Description'].str.strip()\n",
    "\n",
    "df['StockCode'] = df['StockCode'].str.upper()\n",
    "\n",
    "## description that comes back the most per 'StockCode'. if a list comes back, choose first element.\n",
    "stockCode = df.groupby(['StockCode'])['Description'].agg(lambda x: x.mode(dropna=False).iloc[0]).to_frame().rename(columns={'Description':'Description2'}) #.agg(pd.Series.mode).rename('Description2')\n",
    "\n",
    "print('Empty descriptions before fix:',len(df[df['Description'].isnull()]))\n",
    "df = df.merge(stockCode,how='left',left_on='StockCode',right_on='StockCode').drop('Description',axis=1).rename(columns={'Description2':'Description'})\n",
    "df['Description'] = df['Description'].astype(str).apply(lambda x: np.nan if len(x)==0 else x)\n",
    "print('Empty descriptions after fix:',len(df[df['Description'].isnull()]))\n",
    "# df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "531495\n",
      "396825\n"
     ]
    }
   ],
   "source": [
    "print(len(df))\n",
    "df = df[~df['CustomerID'].isnull()]\n",
    "print(len(df))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. RFM Segmentation - Recency, Frequency and Monetary Value Analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>When Will Your Customers Purchase Next?</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>CustomerID</th>\n",
       "      <th>FirstInvoice</th>\n",
       "      <th>LastInvoice</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>DaysBetweenFirstLastInv</th>\n",
       "      <th>lastDate</th>\n",
       "      <th>days_as_customer</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>AvgDaysBetweenPurchases</th>\n",
       "      <th>NextPurchaseInDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>13173</td>\n",
       "      <td>2010-12-07 14:30:00</td>\n",
       "      <td>2011-12-05 15:51:00</td>\n",
       "      <td>2</td>\n",
       "      <td>353.65</td>\n",
       "      <td>363.0</td>\n",
       "      <td>2011-12-09 12:50:00</td>\n",
       "      <td>366.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>363.0</td>\n",
       "      <td>360.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14594</td>\n",
       "      <td>2010-12-01 13:54:00</td>\n",
       "      <td>2011-12-02 13:28:00</td>\n",
       "      <td>2</td>\n",
       "      <td>379.83</td>\n",
       "      <td>365.0</td>\n",
       "      <td>2011-12-09 12:50:00</td>\n",
       "      <td>372.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>365.0</td>\n",
       "      <td>359.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14865</td>\n",
       "      <td>2010-12-02 18:23:00</td>\n",
       "      <td>2011-12-02 16:25:00</td>\n",
       "      <td>2</td>\n",
       "      <td>52.20</td>\n",
       "      <td>364.0</td>\n",
       "      <td>2011-12-09 12:50:00</td>\n",
       "      <td>371.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>364.0</td>\n",
       "      <td>358.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13255</td>\n",
       "      <td>2010-12-01 11:58:00</td>\n",
       "      <td>2011-11-30 13:14:00</td>\n",
       "      <td>2</td>\n",
       "      <td>399.51</td>\n",
       "      <td>364.0</td>\n",
       "      <td>2011-12-09 12:50:00</td>\n",
       "      <td>373.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>364.0</td>\n",
       "      <td>356.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15822</td>\n",
       "      <td>2010-12-15 13:38:00</td>\n",
       "      <td>2011-12-06 14:50:00</td>\n",
       "      <td>2</td>\n",
       "      <td>885.47</td>\n",
       "      <td>356.0</td>\n",
       "      <td>2011-12-09 12:50:00</td>\n",
       "      <td>358.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>356.0</td>\n",
       "      <td>354.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df1 = df.groupby(['CustomerID']).agg({'InvoiceDate':[min,max],'InvoiceNo':pd.Series.nunique,'TotalPrice':np.sum}).reset_index()\n",
    "df1.columns = [' '.join(col).strip() for col in df1.columns.values]\n",
    "df1 = df1.rename(columns={'InvoiceDate min':'FirstInvoice','InvoiceDate max':'LastInvoice','InvoiceNo nunique':'NumInvoice','TotalPrice sum':'TotalSales'})\n",
    "## DaysBetweenFirstLastInv\n",
    "df1['DaysBetweenFirstLastInv'] = (df1['LastInvoice']-df1['FirstInvoice']).astype('timedelta64[D]') # 1 +\n",
    "## days_as_customer\n",
    "df1['lastDate'] = df['InvoiceDate'].max()\n",
    "df1['days_as_customer'] = (df1['lastDate']-df1['FirstInvoice']).astype('timedelta64[D]') #1 + \n",
    "## days_since_purchase\n",
    "df1['days_since_purchase'] = (df1['lastDate']-df1['LastInvoice']).astype('timedelta64[D]') #1 + \n",
    "## Average days between purchases:\n",
    "df1['AvgDaysBetweenPurchases'] = df1['DaysBetweenFirstLastInv'] / (df1['NumInvoice']-1)\n",
    "## Days Above Average: 'days_since_purchase' -  'AvgDaysBetweenPurchases' or 'NextPurchaseInDays': 'AvgDaysBetweenPurchases' - 'days_since_purchase'\n",
    "df1['NextPurchaseInDays'] = df1['AvgDaysBetweenPurchases'] - df1['days_since_purchase']\n",
    "\n",
    "df1 = df1.sort_values('NextPurchaseInDays',ascending=False)\n",
    "\n",
    "display_side_by_side(df1.head(),titles=['When Will Your Customers Purchase Next?'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>326</th>\n",
       "      <td>12748</td>\n",
       "      <td>0.0</td>\n",
       "      <td>210</td>\n",
       "      <td>33715.73</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1880</th>\n",
       "      <td>14911</td>\n",
       "      <td>0.0</td>\n",
       "      <td>201</td>\n",
       "      <td>143825.06</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4011</th>\n",
       "      <td>17841</td>\n",
       "      <td>1.0</td>\n",
       "      <td>124</td>\n",
       "      <td>40991.57</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     CustomerID  days_since_purchase  NumInvoice  TotalSales segment\n",
       "326       12748                  0.0         210    33715.73   1-1-1\n",
       "1880      14911                  0.0         201   143825.06   1-1-1\n",
       "4011      17841                  1.0         124    40991.57   1-1-1"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.drop(['FirstInvoice','LastInvoice','DaysBetweenFirstLastInv','days_as_customer','lastDate','AvgDaysBetweenPurchases','NextPurchaseInDays'],axis=1, inplace=True)\n",
    "\n",
    "rfm = df1[['CustomerID','days_since_purchase','NumInvoice','TotalSales']].copy().sort_values('NumInvoice', ascending=False)\n",
    "\n",
    "rfm['days_since_purchase_Q'] = pd.qcut(rfm['days_since_purchase'].rank(method='first'), 4, labels=[1,2,3,4])\n",
    "rfm['NumInvoice_Q'] = pd.qcut(rfm['NumInvoice'].rank(method='first'), 4, labels=[4,3,2,1])\n",
    "rfm['TotalPrice_Q'] = pd.qcut(rfm['TotalSales'].rank(method='first'), 4, labels=[4,3,2,1])\n",
    "\n",
    "rfm['segment'] = rfm['days_since_purchase_Q'].astype(str) + \"-\" + rfm['NumInvoice_Q'].astype(str) + \"-\" + rfm['TotalPrice_Q'].astype(str)\n",
    "rfm2 = rfm.copy()\n",
    "rfm.drop(['days_since_purchase_Q','NumInvoice_Q','TotalPrice_Q'], axis=1,inplace=True)\n",
    "\n",
    "rfm.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>Best_Customers - count= 500</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>CustomerID</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>14646</td>\n",
       "      <td>1.0</td>\n",
       "      <td>73</td>\n",
       "      <td>279138.02</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18102</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60</td>\n",
       "      <td>259657.30</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17450</td>\n",
       "      <td>7.0</td>\n",
       "      <td>46</td>\n",
       "      <td>194550.79</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14911</td>\n",
       "      <td>0.0</td>\n",
       "      <td>201</td>\n",
       "      <td>143825.06</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14156</td>\n",
       "      <td>9.0</td>\n",
       "      <td>55</td>\n",
       "      <td>117379.63</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th><th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>High_spending_New_Customers - count= 42</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>CustomerID</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>16446</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>168472.50</td>\n",
       "      <td>1-3-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15195</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "      <td>3861.00</td>\n",
       "      <td>1-3-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13868</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2</td>\n",
       "      <td>3192.54</td>\n",
       "      <td>1-3-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15123</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2199.31</td>\n",
       "      <td>1-3-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15531</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2007.44</td>\n",
       "      <td>1-3-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th><th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>Lowest_Spending_Active_Loyal_Customers - count= 8</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>CustomerID</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>17988</td>\n",
       "      <td>11.0</td>\n",
       "      <td>7</td>\n",
       "      <td>201.12</td>\n",
       "      <td>1-1-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16500</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6</td>\n",
       "      <td>400.86</td>\n",
       "      <td>1-1-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16668</td>\n",
       "      <td>14.0</td>\n",
       "      <td>6</td>\n",
       "      <td>306.72</td>\n",
       "      <td>1-1-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16915</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6</td>\n",
       "      <td>621.68</td>\n",
       "      <td>1-1-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16892</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>525.54</td>\n",
       "      <td>1-1-3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th><th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>Churned_Best_Customers - count= 81</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>CustomerID</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>15749</td>\n",
       "      <td>234.0</td>\n",
       "      <td>3</td>\n",
       "      <td>44534.30</td>\n",
       "      <td>4-2-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15098</td>\n",
       "      <td>181.0</td>\n",
       "      <td>3</td>\n",
       "      <td>39916.50</td>\n",
       "      <td>4-2-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13093</td>\n",
       "      <td>275.0</td>\n",
       "      <td>8</td>\n",
       "      <td>7832.47</td>\n",
       "      <td>4-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12980</td>\n",
       "      <td>157.0</td>\n",
       "      <td>9</td>\n",
       "      <td>7374.90</td>\n",
       "      <td>4-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16553</td>\n",
       "      <td>162.0</td>\n",
       "      <td>12</td>\n",
       "      <td>5719.82</td>\n",
       "      <td>4-1-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Best_Customers = rfm[rfm['segment'] == '1-1-1'].sort_values('TotalSales',ascending=False)\n",
    "\n",
    "listy = ['1-4-1','1-4-2','1-3-1','1-3-2']\n",
    "High_spending_New_Customers = rfm[rfm['segment'].isin(listy)].sort_values('TotalSales',ascending=False)\n",
    "\n",
    "listy = ['1-1-3','1-1-4']\n",
    "Lowest_Spending_Active_Loyal_Customers = rfm[rfm['segment'].isin(listy)].sort_values('NumInvoice',ascending=False)\n",
    "\n",
    "listy = ['4-1-1','4-1-2','4-2-1','4-2-2']\n",
    "Churned_Best_Customers = rfm[rfm['segment'].isin(listy)].sort_values('TotalSales',ascending=False)\n",
    "\n",
    "display_side_by_side(Best_Customers.head(), High_spending_New_Customers.head(),\n",
    "                     Lowest_Spending_Active_Loyal_Customers.head(),Churned_Best_Customers.head(),\n",
    "                     titles=['Best_Customers - count= {}'.format(len(Best_Customers)),'High_spending_New_Customers - count= {}'.format(len(High_spending_New_Customers)),\n",
    "                             'Lowest_Spending_Active_Loyal_Customers - count= {}'.format(len(Lowest_Spending_Active_Loyal_Customers)),'Churned_Best_Customers - count= {}'.format(len(Churned_Best_Customers))])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 263,
       "width": 1178
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "listo = ['days_since_purchase','NumInvoice','TotalSales']\n",
    "\n",
    "fig,ax = plt.subplots(1,3,figsize=(20,4))\n",
    "for i in range(len(listo)):\n",
    "    sns.distplot(Best_Customers[listo[i]], hist=True, ax=ax[i]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Let's merge back to keep our best customers (segment 1-1-1):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "167119\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>InvoiceNo</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>UnitPrice</th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>Country</th>\n",
       "      <th>TotalPrice</th>\n",
       "      <th>day</th>\n",
       "      <th>day_name</th>\n",
       "      <th>day_of_year</th>\n",
       "      <th>week</th>\n",
       "      <th>hour</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "      <th>weekday</th>\n",
       "      <th>weekend</th>\n",
       "      <th>month_year</th>\n",
       "      <th>Description</th>\n",
       "      <th>days_since_purchase</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>536370</td>\n",
       "      <td>22728</td>\n",
       "      <td>24</td>\n",
       "      <td>2010-12-01 08:45:00</td>\n",
       "      <td>3.75</td>\n",
       "      <td>12583</td>\n",
       "      <td>France</td>\n",
       "      <td>90.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "      <td>ALARM CLOCK BAKELIKE PINK</td>\n",
       "      <td>2.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>6705.38</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>536370</td>\n",
       "      <td>22727</td>\n",
       "      <td>24</td>\n",
       "      <td>2010-12-01 08:45:00</td>\n",
       "      <td>3.75</td>\n",
       "      <td>12583</td>\n",
       "      <td>France</td>\n",
       "      <td>90.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "      <td>ALARM CLOCK BAKELIKE RED</td>\n",
       "      <td>2.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>6705.38</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>536370</td>\n",
       "      <td>22726</td>\n",
       "      <td>12</td>\n",
       "      <td>2010-12-01 08:45:00</td>\n",
       "      <td>3.75</td>\n",
       "      <td>12583</td>\n",
       "      <td>France</td>\n",
       "      <td>45.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Wednesday</td>\n",
       "      <td>335</td>\n",
       "      <td>48</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2010</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-12</td>\n",
       "      <td>ALARM CLOCK BAKELIKE GREEN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>6705.38</td>\n",
       "      <td>1-1-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   InvoiceNo StockCode  Quantity         InvoiceDate  UnitPrice CustomerID  \\\n",
       "26    536370     22728        24 2010-12-01 08:45:00       3.75      12583   \n",
       "27    536370     22727        24 2010-12-01 08:45:00       3.75      12583   \n",
       "28    536370     22726        12 2010-12-01 08:45:00       3.75      12583   \n",
       "\n",
       "   Country  TotalPrice  day   day_name  day_of_year  week  hour  month  year  \\\n",
       "26  France        90.0    1  Wednesday          335    48     8     12  2010   \n",
       "27  France        90.0    1  Wednesday          335    48     8     12  2010   \n",
       "28  France        45.0    1  Wednesday          335    48     8     12  2010   \n",
       "\n",
       "    weekday  weekend month_year                 Description  \\\n",
       "26        2        0    2010-12   ALARM CLOCK BAKELIKE PINK   \n",
       "27        2        0    2010-12    ALARM CLOCK BAKELIKE RED   \n",
       "28        2        0    2010-12  ALARM CLOCK BAKELIKE GREEN   \n",
       "\n",
       "    days_since_purchase  NumInvoice  TotalSales segment  \n",
       "26                  2.0        15.0     6705.38   1-1-1  \n",
       "27                  2.0        15.0     6705.38   1-1-1  \n",
       "28                  2.0        15.0     6705.38   1-1-1  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bestCustomer = df.merge(Best_Customers,how='left',left_on='CustomerID',right_on='CustomerID')\n",
    "bestCustomer = bestCustomer[~bestCustomer['segment'].isnull()]\n",
    "print(len(bestCustomer))\n",
    "bestCustomer.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Singular Value Decomposition (SVD) to recommend purchases using our best customers:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Here, we do not have a rating column so let's create one by looking at how many times a product has been purchased (# of invoices) divided by the total number of invoices for each customer and create bins for rating between 1 to 5.\n",
    "(this is why I keep the best customers only to be more meaningful: more invoices, customers for a while, etc...)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "83493 1 5\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>Description</th>\n",
       "      <th>InvoiceNo</th>\n",
       "      <th>NumInvoice</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12347</td>\n",
       "      <td>3D DOG PICTURE PLAYING CARDS</td>\n",
       "      <td>5</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12347</td>\n",
       "      <td>3D SHEET OF CAT STICKERS</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12347</td>\n",
       "      <td>3D SHEET OF DOG STICKERS</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  CustomerID                   Description  InvoiceNo  NumInvoice  rating\n",
       "0      12347  3D DOG PICTURE PLAYING CARDS          5         7.0       4\n",
       "1      12347      3D SHEET OF CAT STICKERS          1         7.0       1\n",
       "2      12347      3D SHEET OF DOG STICKERS          1         7.0       1"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "freq = bestCustomer.groupby(['CustomerID','Description']).agg({'InvoiceNo':pd.Series.nunique,'NumInvoice':np.unique}).reset_index()\n",
    "freq['rating'] = freq['InvoiceNo'] / freq['NumInvoice']\n",
    "\n",
    "freq['rating'] = pd.cut(freq['rating'], bins=[0,0.2,0.4,0.6,0.8,1], labels=[1,2,3,4,5])\n",
    "freq['rating'] = freq['rating'].astype(int)\n",
    "\n",
    "print(len(freq),min(freq['rating']),max(freq['rating']))\n",
    "freq.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Let's create also a unique column of descriptions of purchases to use below with the SVD:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3408\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3D DOG PICTURE PLAYING CARDS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3D SHEET OF CAT STICKERS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3D SHEET OF DOG STICKERS</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    Description\n",
       "0  3D DOG PICTURE PLAYING CARDS\n",
       "1      3D SHEET OF CAT STICKERS\n",
       "2      3D SHEET OF DOG STICKERS"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "desc = pd.DataFrame()\n",
    "desc['Description'] = freq['Description'].unique()\n",
    "print(len(desc))\n",
    "desc.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import packages used for the SVD:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import Reader, Dataset, SVD\n",
    "from surprise.model_selection.validation import cross_validate\n",
    "from surprise.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The Dataset module in Surprise provides different methods for loading data from files, Pandas DataFrames, or built-in datasets:**\n",
    "\n",
    "Dataset.load_builtin()\n",
    "Dataset.load_from_file()\n",
    "Dataset.load_from_df()\n",
    "\n",
    "The **Reader class** in Surprise is to parse a file containing users, items, and users’ ratings on items. The default format is that each rating is stored in a separate line in the following order separated by space: user item rating\n",
    "This order and the separator are configurable using the following parameters:\n",
    "\n",
    "- **line_format** is a string like “item user rating” to indicate the order of the data with field names separated by a space\n",
    "- **sep** is used to specify separator between fields, such as space, ‘,’, etc.\n",
    "- **rating_scale** is to specify the rating scale. The default value is (1, 5)\n",
    "- **skip_lines** is to indicate the number of lines to skip at the beginning of the file and the default is 0\n",
    "\n",
    "let's load the dataframe and create a SVD model instance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "reader = Reader(rating_scale=[1,5])\n",
    "\n",
    "data = Dataset.load_from_df(freq[['CustomerID','Description','rating']], reader)\n",
    "svd = SVD()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hypertuning of SVD:\n",
    "(trying multiple combinations until you fin the best/optimised parameters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.4882955892277372\n",
      "{'n_factors': 22, 'n_epochs': 38, 'lr_all': 0.004, 'reg_all': 0.1}\n",
      "CPU times: user 6.78 s, sys: 35.4 ms, total: 6.81 s\n",
      "Wall time: 6.82 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "param_grid = {\n",
    "              \"n_factors\":[22],\n",
    "              \"n_epochs\": [38],\n",
    "              \"lr_all\": [0.004],\n",
    "              \"reg_all\": [0.1]\n",
    "                }\n",
    "gs = GridSearchCV(SVD, param_grid, measures=[\"rmse\", \"mae\"], cv=3)\n",
    "\n",
    "gs.fit(data)\n",
    "\n",
    "print(gs.best_score[\"rmse\"])\n",
    "print(gs.best_params[\"rmse\"])\n",
    "\n",
    "params = gs.best_params['rmse']\n",
    "svdtuned = SVD(n_factors=params['n_factors'], \n",
    "               n_epochs=params['n_epochs'],\n",
    "               lr_all=params['lr_all'], \n",
    "               reg_all=params['reg_all'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Run 5-fold cross-validation and print results\n",
    "# cross_validate(svdtuned, data, measures=['RMSE', 'MAE'], cv=5, verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Once the model has been evaluated to our satisfaction, then we can re-train the model using the entire training dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<surprise.prediction_algorithms.matrix_factorization.SVD at 0x13b7d7a60>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainset = data.build_full_trainset()\n",
    "svdtuned.fit(trainset)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Recommending Products:\n",
    "\n",
    "After the recommendation model has been trained appropriately, it can be used for prediction.\n",
    "For example, given a customer, we can use the trained model to predict products in which the customer may be interested."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3408\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>Product recommendations for customer #16360</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Description</th>\n",
       "      <th>Estimate_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>POPCORN HOLDER</td>\n",
       "      <td>2.465244</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>WHITE HANGING HEART T-LIGHT HOLDER</td>\n",
       "      <td>2.266821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>ALARM CLOCK BAKELIKE GREEN</td>\n",
       "      <td>2.143308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>JUMBO BAG RED RETROSPOT</td>\n",
       "      <td>2.137228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>TABLE LAMP WHITE SHADE WOOD BASE</td>\n",
       "      <td>2.128942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>LUNCH BAG SUKI DESIGN</td>\n",
       "      <td>2.093131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>ASSORTED COLOUR BIRD ORNAMENT</td>\n",
       "      <td>2.079600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>WOOD S/3 CABINET ANT WHITE FINISH</td>\n",
       "      <td>2.056717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>VICTORIAN GLASS HANGING T-LIGHT</td>\n",
       "      <td>2.049609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CHILLI LIGHTS</td>\n",
       "      <td>2.037471</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## For one customer:\n",
    "customer = '16360' #'15993' #'14911'\n",
    "\n",
    "descrip = desc.copy()\n",
    "\n",
    "descrip['Estimate_Score'] = descrip['Description'].apply(lambda x: svdtuned.predict(customer, x).est)\n",
    "descrip = descrip.sort_values(by=['Estimate_Score'], ascending=False)\n",
    "print(len(descrip))\n",
    "display_side_by_side(descrip.head(10),titles=['Product recommendations for customer #{}'.format(customer)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ## For all customers:\n",
    "# list_df = []\n",
    "# for i in freq['CustomerID'].unique():\n",
    "#     descrip = desc.copy()\n",
    "#     descrip['Estimate_Score'] = descrip['Description'].apply(lambda x: svdtuned.predict(i, x).est)\n",
    "#     descrip = descrip.sort_values(by=['Estimate_Score'], ascending=False)\n",
    "#     descrip['CustomerID'] = i\n",
    "#     list_df.append(descrip.head(10))\n",
    "\n",
    "# new_df = pd.concat(list_df).reset_index(drop=True)\n",
    "# new_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Market-Basket Analysis also known as 'people who bought this also bought...':"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "decomposed matrix shape: (3602, 100)\n",
      "correlation matrix shape: (3602, 3602)\n"
     ]
    }
   ],
   "source": [
    "import sklearn\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "\n",
    "## shape the 'rating' dataset into a matrix\n",
    "basket = df.groupby(['InvoiceNo', 'Description']).agg({'Quantity':np.sum}).unstack().reset_index().fillna(0).set_index('InvoiceNo')\n",
    "## replace values bigger than 1 by one:\n",
    "basket[basket>1]=1\n",
    "## fix columns name:\n",
    "basket.index.name = None\n",
    "basket.columns = [col[1] for col in basket.columns]\n",
    "## Transpose\n",
    "basket = basket.T\n",
    "# print(basket.head())\n",
    "\n",
    "## Dimensionality reduction using truncated SVD (aka LSA). \n",
    "## This transformer performs linear dimensionality reduction by means of truncated singular value decomposition:\n",
    "TruncSVD = TruncatedSVD(n_components=100)\n",
    "decomposed = TruncSVD.fit_transform(basket)\n",
    "print('decomposed matrix shape:',decomposed.shape)\n",
    "\n",
    "## Correlation matrix:\n",
    "correlation = np.corrcoef(decomposed)\n",
    "print('correlation matrix shape:',correlation.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import random\n",
    "\n",
    "# ## Choose a random product:\n",
    "# foo = freq['Description'].unique()\n",
    "# product = random.choice(foo)\n",
    "# print(product)\n",
    "\n",
    "product = 'GARDENERS KNEELING PAD CUP OF TEA'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index of product: 1264\n",
      "Number of correlations of product: (3602,)\n"
     ]
    }
   ],
   "source": [
    "list_descriptions = list(basket.index)\n",
    "product_index = list_descriptions.index(product)\n",
    "print('Index of product:',product_index)\n",
    "\n",
    "correlation_product_index = correlation[product_index]\n",
    "print('Number of correlations of product:',correlation_product_index.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of related products above threshold: 2\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<th style=\"text-align:center\"><td style=\"vertical-align:top\"><h2>People who bought GARDENERS KNEELING PAD CUP OF TEA, also bought:</h2><table style=\"display:inline\" border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Description</th>\n",
       "      <th>Correlation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>GARDENERS KNEELING PAD KEEP CALM</td>\n",
       "      <td>0.986484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>KNEELING MAT HOUSEWORK  DESIGN</td>\n",
       "      <td>0.927839</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table style=\"display:inline\"></td></th>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "threshold = 0.7\n",
    "Recommend = pd.DataFrame(list([(x,y) for x,y in zip(basket.index[correlation_product_index > threshold],\n",
    "                                                    correlation_product_index[correlation_product_index > threshold])]),columns=['Description','Correlation'])\n",
    "Recommend = Recommend[Recommend['Description'] != product].sort_values('Correlation',ascending=False)\n",
    "print('Number of related products above threshold:',len(Recommend))\n",
    "display_side_by_side(Recommend.head(),titles=['People who bought {}, also bought:'.format(product)])"
   ]
  },
  {
   "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}