chapter_3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chapter 3 Feature Engineering for Marketing Data\n",
    "\n",
    "In the previous chapter, we learned how to preprocess and explore our marketing data using Python libraries like Pandas, Matplotlib, and Seaborn. Now, we'll dive into feature engineering, a crucial step in the data science workflow that involves creating new and informative features from existing data to improve the predictive power of our models.\n",
    "\n",
    "Feature engineering is particularly important in marketing analytics, as it helps us capture more complex patterns and relationships in customer behavior and market trends. In this chapter, we'll explore various techniques for creating new features using Python and apply them to our Online Retail dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### 3.1 Recency, Frequency, and Monetary (RFM) Analysis\n",
    "One of the most common feature engineering techniques in marketing is Recency, Frequency, and Monetary (RFM) analysis. RFM analysis segments customers based on three key metrics:\n",
    "\n",
    "- Recency: How recently a customer made a purchase\n",
    "- Frequency: How often a customer makes purchases\n",
    "- Monetary: How much a customer spends\n",
    "\n",
    "Let's calculate RFM metrics for our Online Retail dataset using Python:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "      <th>Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>489434</td>\n",
       "      <td>85048</td>\n",
       "      <td>15CM CHRISTMAS GLASS BALL 20 LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.95</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>83.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323P</td>\n",
       "      <td>PINK CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323W</td>\n",
       "      <td>WHITE CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>489434</td>\n",
       "      <td>22041</td>\n",
       "      <td>RECORD FRAME 7\" SINGLE SIZE</td>\n",
       "      <td>48</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>2.10</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>100.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>489434</td>\n",
       "      <td>21232</td>\n",
       "      <td>STRAWBERRY CERAMIC TRINKET BOX</td>\n",
       "      <td>24</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>1.25</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Invoice StockCode                          Description  Quantity  \\\n",
       "0  489434     85048  15CM CHRISTMAS GLASS BALL 20 LIGHTS        12   \n",
       "1  489434    79323P                   PINK CHERRY LIGHTS        12   \n",
       "2  489434    79323W                  WHITE CHERRY LIGHTS        12   \n",
       "3  489434     22041         RECORD FRAME 7\" SINGLE SIZE         48   \n",
       "4  489434     21232       STRAWBERRY CERAMIC TRINKET BOX        24   \n",
       "\n",
       "          InvoiceDate  Price  Customer ID         Country  Amount  \n",
       "0 2009-12-01 07:45:00   6.95      13085.0  United Kingdom    83.4  \n",
       "1 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0  \n",
       "2 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0  \n",
       "3 2009-12-01 07:45:00   2.10      13085.0  United Kingdom   100.8  \n",
       "4 2009-12-01 07:45:00   1.25      13085.0  United Kingdom    30.0  "
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "# read the csv file into a pandas dataframe \n",
    "df = pd.read_csv('data\\online_retail.csv')\n",
    "df['Amount'] = df['Quantity'] * df['Price']\n",
    "df['InvoiceDate'] = pd.to_datetime(df['InvoiceDate'])\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Customer ID\n",
       "12346.0   2011-01-18 10:17:00\n",
       "12347.0   2011-12-07 15:52:00\n",
       "12348.0   2011-09-25 13:13:00\n",
       "12349.0   2011-11-21 09:51:00\n",
       "12350.0   2011-02-02 16:01:00\n",
       "                  ...        \n",
       "18283.0   2011-12-06 12:02:00\n",
       "18284.0   2010-10-06 12:31:00\n",
       "18285.0   2010-02-17 10:24:00\n",
       "18286.0   2010-08-20 11:57:00\n",
       "18287.0   2011-10-28 09:29:00\n",
       "Name: InvoiceDate, Length: 5942, dtype: datetime64[ns]"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from datetime import datetime\n",
    "\n",
    "# Calculate the maximum InvoiceDate for each customer\n",
    "max_date = df.groupby('Customer ID')['InvoiceDate'].max()\n",
    "max_date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "scoring_date = datetime(2011, 12, 10)\n",
    "\n",
    "# Calculate Recency, Frequency, and Monetary metrics\n",
    "rfm = df.groupby('Customer ID').agg({\n",
    "    'InvoiceDate': lambda x: (scoring_date - x.max()).days,\n",
    "    'Invoice': 'nunique',\n",
    "    'Amount': 'sum'\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rename columns\n",
    "rfm.columns = ['Recency', 'Frequency', 'Monetary']\n",
    "\n",
    "# Assign quintile ranks to each metric\n",
    "rfm['R_rank'] = pd.qcut(rfm['Recency'], 5, labels=False, duplicates='drop')\n",
    "rfm['F_rank'] = pd.qcut(rfm['Frequency'], 5, labels=False, duplicates='drop')\n",
    "rfm['M_rank'] = pd.qcut(rfm['Monetary'], 5, labels=False, duplicates='drop')\n",
    "\n",
    "# Add 1 to the ranks to start ranking from 1 instead of 0\n",
    "rfm['R_rank'] += 1\n",
    "rfm['F_rank'] += 1\n",
    "rfm['M_rank'] += 1\n",
    "\n",
    "# Combine R, F, and M ranks into a single RFM score\n",
    "rfm['RFM_score'] = rfm['R_rank'].astype(str) + rfm['F_rank'].astype(str) + rfm['M_rank'].astype(str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "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>Recency</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "      <th>R_rank</th>\n",
       "      <th>F_rank</th>\n",
       "      <th>M_rank</th>\n",
       "      <th>RFM_score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12346.0</th>\n",
       "      <td>325</td>\n",
       "      <td>17</td>\n",
       "      <td>-64.68</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>5633.32</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>75</td>\n",
       "      <td>5</td>\n",
       "      <td>2019.40</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12349.0</th>\n",
       "      <td>18</td>\n",
       "      <td>5</td>\n",
       "      <td>4404.54</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12350.0</th>\n",
       "      <td>310</td>\n",
       "      <td>1</td>\n",
       "      <td>334.40</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>412</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency  Frequency  Monetary  R_rank  F_rank  M_rank RFM_score\n",
       "Customer ID                                                                \n",
       "12346.0          325         17    -64.68       4       4       1       441\n",
       "12347.0            2          8   5633.32       1       3       5       135\n",
       "12348.0           75          5   2019.40       3       2       4       324\n",
       "12349.0           18          5   4404.54       1       2       5       125\n",
       "12350.0          310          1    334.40       4       1       2       412"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfm.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this code snippet, we first calculate the maximum InvoiceDate for each customer. Then, we calculate the Recency, Frequency, and Monetary metrics using the groupby() and agg() functions. We assign quintile ranks to each metric using pd.qcut() and combine the ranks into a single RFM score.\n",
    "\n",
    "The resulting RFM score can be used to segment customers and tailor marketing strategies accordingly. For example, customers with high RFM scores (e.g., 555) are likely to be high-value customers who have made recent, frequent, and high-value purchases.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By calculating RFM metrics for the Online Retail dataset, we can segment our customers based on their purchasing behavior and create targeted marketing campaigns. For example, we might offer incentives to customers in the lowest recency quintiles (Q1 or Q2) to encourage them to make a purchase. Similarly, we might target customers in the highest monetary quintiles (Q4 or Q5) with high-value products or services.\n",
    "\n",
    "Now that we've calculated RFM metrics for our dataset, let's move on to other feature engineering techniques for marketing data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## 3.2 Time-based Features\n",
    "\n",
    "Time-based features can provide valuable insights into customer behavior over time. For example, we might want to understand how a customer's purchasing behavior changes throughout the year or the impact of a marketing campaign on sales over time.\n",
    "\n",
    "To create time-based features for our Online Retail dataset, we might use techniques such as:\n",
    "\n",
    "* Rolling averages and moving averages to capture trends in customer behavior over time\n",
    "* Lagged features to capture the impact of previous purchases or events on current behavior\n",
    "* Year, month, day, and hour-of-the-day features to capture seasonality or time-specific patterns\n",
    "\n",
    "Let's create time-based features for our Online Retail dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "      <th>Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>489434</td>\n",
       "      <td>85048</td>\n",
       "      <td>15CM CHRISTMAS GLASS BALL 20 LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.95</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>83.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323P</td>\n",
       "      <td>PINK CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323W</td>\n",
       "      <td>WHITE CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>489434</td>\n",
       "      <td>22041</td>\n",
       "      <td>RECORD FRAME 7\" SINGLE SIZE</td>\n",
       "      <td>48</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>2.10</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>100.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>489434</td>\n",
       "      <td>21232</td>\n",
       "      <td>STRAWBERRY CERAMIC TRINKET BOX</td>\n",
       "      <td>24</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>1.25</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Invoice StockCode                          Description  Quantity  \\\n",
       "0  489434     85048  15CM CHRISTMAS GLASS BALL 20 LIGHTS        12   \n",
       "1  489434    79323P                   PINK CHERRY LIGHTS        12   \n",
       "2  489434    79323W                  WHITE CHERRY LIGHTS        12   \n",
       "3  489434     22041         RECORD FRAME 7\" SINGLE SIZE         48   \n",
       "4  489434     21232       STRAWBERRY CERAMIC TRINKET BOX        24   \n",
       "\n",
       "          InvoiceDate  Price  Customer ID         Country  Amount  \n",
       "0 2009-12-01 07:45:00   6.95      13085.0  United Kingdom    83.4  \n",
       "1 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0  \n",
       "2 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0  \n",
       "3 2009-12-01 07:45:00   2.10      13085.0  United Kingdom   100.8  \n",
       "4 2009-12-01 07:45:00   1.25      13085.0  United Kingdom    30.0  "
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "      <th>Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1824</th>\n",
       "      <td>C489592</td>\n",
       "      <td>22301</td>\n",
       "      <td>COFFEE MUG CAT + BIRD DESIGN</td>\n",
       "      <td>-2</td>\n",
       "      <td>2009-12-01 14:19:00</td>\n",
       "      <td>2.55</td>\n",
       "      <td>18287.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>-5.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199795</th>\n",
       "      <td>508581</td>\n",
       "      <td>22243</td>\n",
       "      <td>HOOK, 5 HANGER ,MAGIC TOADSTOOL RED</td>\n",
       "      <td>12</td>\n",
       "      <td>2010-05-17 11:55:00</td>\n",
       "      <td>1.65</td>\n",
       "      <td>18287.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>19.80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199796</th>\n",
       "      <td>508581</td>\n",
       "      <td>22246</td>\n",
       "      <td>GARLAND, MAGIC GARDEN 1.8M</td>\n",
       "      <td>12</td>\n",
       "      <td>2010-05-17 11:55:00</td>\n",
       "      <td>1.95</td>\n",
       "      <td>18287.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>23.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199797</th>\n",
       "      <td>508581</td>\n",
       "      <td>22236</td>\n",
       "      <td>CAKE STAND 3 TIER MAGIC GARDEN</td>\n",
       "      <td>1</td>\n",
       "      <td>2010-05-17 11:55:00</td>\n",
       "      <td>12.75</td>\n",
       "      <td>18287.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>12.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199798</th>\n",
       "      <td>508581</td>\n",
       "      <td>21439</td>\n",
       "      <td>BASKET OF TOADSTOOLS</td>\n",
       "      <td>12</td>\n",
       "      <td>2010-05-17 11:55:00</td>\n",
       "      <td>1.25</td>\n",
       "      <td>18287.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>15.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Invoice StockCode                          Description  Quantity  \\\n",
       "1824    C489592     22301         COFFEE MUG CAT + BIRD DESIGN        -2   \n",
       "199795   508581     22243  HOOK, 5 HANGER ,MAGIC TOADSTOOL RED        12   \n",
       "199796   508581     22246           GARLAND, MAGIC GARDEN 1.8M        12   \n",
       "199797   508581     22236       CAKE STAND 3 TIER MAGIC GARDEN         1   \n",
       "199798   508581     21439                 BASKET OF TOADSTOOLS        12   \n",
       "\n",
       "               InvoiceDate  Price  Customer ID         Country  Amount  \n",
       "1824   2009-12-01 14:19:00   2.55      18287.0  United Kingdom   -5.10  \n",
       "199795 2010-05-17 11:55:00   1.65      18287.0  United Kingdom   19.80  \n",
       "199796 2010-05-17 11:55:00   1.95      18287.0  United Kingdom   23.40  \n",
       "199797 2010-05-17 11:55:00  12.75      18287.0  United Kingdom   12.75  \n",
       "199798 2010-05-17 11:55:00   1.25      18287.0  United Kingdom   15.00  "
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[df['Customer ID'] == 18287.0].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "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></th>\n",
       "      <th>Amount</th>\n",
       "      <th>Invoice</th>\n",
       "      <th>Quantity</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">12346.0</th>\n",
       "      <th>2009-12-14</th>\n",
       "      <td>90.0</td>\n",
       "      <td>3</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-15</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-16</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-17</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-18</th>\n",
       "      <td>23.5</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Amount  Invoice  Quantity\n",
       "Customer ID InvoiceDate                           \n",
       "12346.0     2009-12-14     90.0        3        20\n",
       "            2009-12-15      0.0        0         0\n",
       "            2009-12-16      0.0        0         0\n",
       "            2009-12-17      0.0        0         0\n",
       "            2009-12-18     23.5        2         6"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Set the index of the dataFrame to InvoiceDate\n",
    "df_tbf = df.copy()\n",
    "df_tbf.set_index('InvoiceDate', inplace=True)\n",
    "\n",
    "# create a sum of sales for each customer for each day, number of unique invoices, and total sales\n",
    "daily_sales = df_tbf.groupby('Customer ID').resample('D').agg({'Amount': 'sum', 'Invoice': 'nunique', 'Quantity': 'sum'}) \n",
    "daily_sales.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "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>Customer ID</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>RollingAvgSales_7days</th>\n",
       "      <th>RollingAvgInvoices_7days</th>\n",
       "      <th>RollingAvgQuantity_7days</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1644576</th>\n",
       "      <td>18287.0</td>\n",
       "      <td>2011-10-24</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1644577</th>\n",
       "      <td>18287.0</td>\n",
       "      <td>2011-10-25</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1644578</th>\n",
       "      <td>18287.0</td>\n",
       "      <td>2011-10-26</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1644579</th>\n",
       "      <td>18287.0</td>\n",
       "      <td>2011-10-27</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1644580</th>\n",
       "      <td>18287.0</td>\n",
       "      <td>2011-10-28</td>\n",
       "      <td>10.097143</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>15.428571</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Customer ID InvoiceDate  RollingAvgSales_7days  \\\n",
       "1644576      18287.0  2011-10-24               0.000000   \n",
       "1644577      18287.0  2011-10-25               0.000000   \n",
       "1644578      18287.0  2011-10-26               0.000000   \n",
       "1644579      18287.0  2011-10-27               0.000000   \n",
       "1644580      18287.0  2011-10-28              10.097143   \n",
       "\n",
       "         RollingAvgInvoices_7days  RollingAvgQuantity_7days  \n",
       "1644576                  0.000000                  0.000000  \n",
       "1644577                  0.000000                  0.000000  \n",
       "1644578                  0.000000                  0.000000  \n",
       "1644579                  0.000000                  0.000000  \n",
       "1644580                  0.142857                 15.428571  "
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create a rolling average of sales for each customer over the last 7 days\n",
    "rolling_avg = daily_sales.groupby('Customer ID')[['Amount',\"Invoice\",\"Quantity\"]].rolling(window=7).mean()\n",
    "rolling_avg = rolling_avg.reset_index(drop=True, level=0).reset_index()\n",
    "# rename the column to RollingAvg\n",
    "rolling_avg = rolling_avg.rename(columns={'Amount': 'RollingAvgSales_7days', \"Invoice\": \"RollingAvgInvoices_7days\", \"Quantity\": \"RollingAvgQuantity_7days\"})\n",
    "rolling_avg.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create lagged features for each customer's total sales and number of orders (invoices) for daily_sales\n",
    "daily_sales['Amount_lag1'] = daily_sales.groupby('Customer ID')['Amount'].shift(1) \n",
    "daily_sales['Amount_lag2'] = daily_sales.groupby('Customer ID')['Amount'].shift(2)\n",
    "daily_sales['InvoiceNo_lag1'] = daily_sales.groupby('Customer ID')['Invoice'].shift(1)\n",
    "daily_sales['InvoiceNo_lag2'] = daily_sales.groupby('Customer ID')['Invoice'].shift(2)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create year, month, day, hour-of-the-day, and Day-of-the-week features from the InvoiceDate index in daily_sales\n",
    "daily_sales['InvoiceYear'] = daily_sales.index.get_level_values('InvoiceDate').year\n",
    "daily_sales['InvoiceMonth'] = daily_sales.index.get_level_values('InvoiceDate').month\n",
    "daily_sales['InvoiceDay'] = daily_sales.index.get_level_values('InvoiceDate').day\n",
    "daily_sales['InvoiceHour'] = daily_sales.index.get_level_values('InvoiceDate').hour\n",
    "daily_sales['InvoiceDayOfWeek'] = daily_sales.index.get_level_values('InvoiceDate').dayofweek\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "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></th>\n",
       "      <th>Amount</th>\n",
       "      <th>Invoice</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>Amount_lag1</th>\n",
       "      <th>Amount_lag2</th>\n",
       "      <th>InvoiceNo_lag1</th>\n",
       "      <th>InvoiceNo_lag2</th>\n",
       "      <th>InvoiceYear</th>\n",
       "      <th>InvoiceMonth</th>\n",
       "      <th>InvoiceDay</th>\n",
       "      <th>InvoiceHour</th>\n",
       "      <th>InvoiceDayOfWeek</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">12346.0</th>\n",
       "      <th>2009-12-14</th>\n",
       "      <td>90.0</td>\n",
       "      <td>3</td>\n",
       "      <td>20</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-15</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-16</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-17</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>17</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-12-18</th>\n",
       "      <td>23.5</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>18</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Amount  Invoice  Quantity  Amount_lag1  Amount_lag2  \\\n",
       "Customer ID InvoiceDate                                                        \n",
       "12346.0     2009-12-14     90.0        3        20          NaN          NaN   \n",
       "            2009-12-15      0.0        0         0         90.0          NaN   \n",
       "            2009-12-16      0.0        0         0          0.0         90.0   \n",
       "            2009-12-17      0.0        0         0          0.0          0.0   \n",
       "            2009-12-18     23.5        2         6          0.0          0.0   \n",
       "\n",
       "                         InvoiceNo_lag1  InvoiceNo_lag2  InvoiceYear  \\\n",
       "Customer ID InvoiceDate                                                \n",
       "12346.0     2009-12-14              NaN             NaN         2009   \n",
       "            2009-12-15              3.0             NaN         2009   \n",
       "            2009-12-16              0.0             3.0         2009   \n",
       "            2009-12-17              0.0             0.0         2009   \n",
       "            2009-12-18              0.0             0.0         2009   \n",
       "\n",
       "                         InvoiceMonth  InvoiceDay  InvoiceHour  \\\n",
       "Customer ID InvoiceDate                                          \n",
       "12346.0     2009-12-14             12          14            0   \n",
       "            2009-12-15             12          15            0   \n",
       "            2009-12-16             12          16            0   \n",
       "            2009-12-17             12          17            0   \n",
       "            2009-12-18             12          18            0   \n",
       "\n",
       "                         InvoiceDayOfWeek  \n",
       "Customer ID InvoiceDate                    \n",
       "12346.0     2009-12-14                  0  \n",
       "            2009-12-15                  1  \n",
       "            2009-12-16                  2  \n",
       "            2009-12-17                  3  \n",
       "            2009-12-18                  4  "
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "daily_sales.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this example, we demonstrate how to extract time-based features from the InvoiceDate column and the methods for doing so. This exercise aims to enhance our understanding of the underlying concepts and fundamental calculation logic. We will show how to utilize specialized Python libraries to generate these features for large datasets."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.3 Holiday-related Features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To capture holiday-related features, we'll use the 'holidays' Python package. This package allows us to easily identify holidays for a given country or region.\n",
    "\n",
    "Let's install the 'holidays' package using pip:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install holidays"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once installed, we can use the holidays package to create holiday-related features for the Online Retail dataset. As the country in the dataset is United Kingdom, we will create the UK holiday-related features. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import holidays\n",
    "\n",
    "# Create a holiday object for the United States\n",
    "uk_holidays = holidays.UK()  # this is a dict\n",
    "\n",
    "# Create a feature indicating if a transaction occurred on a holiday\n",
    "df['Holiday'] = df['InvoiceDate'].apply(lambda x: x.date() in uk_holidays)\n",
    "\n",
    "# Create a feature indicating the name of the holiday (if applicable)\n",
    "df['HolidayName'] = df['InvoiceDate'].apply(lambda x: uk_holidays.get(x.date()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Holiday</th>\n",
       "      <th>HolidayName</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>489434</td>\n",
       "      <td>85048</td>\n",
       "      <td>15CM CHRISTMAS GLASS BALL 20 LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.95</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>83.4</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323P</td>\n",
       "      <td>PINK CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323W</td>\n",
       "      <td>WHITE CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>489434</td>\n",
       "      <td>22041</td>\n",
       "      <td>RECORD FRAME 7\" SINGLE SIZE</td>\n",
       "      <td>48</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>2.10</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>100.8</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>489434</td>\n",
       "      <td>21232</td>\n",
       "      <td>STRAWBERRY CERAMIC TRINKET BOX</td>\n",
       "      <td>24</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>1.25</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>30.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Invoice StockCode                          Description  Quantity  \\\n",
       "0  489434     85048  15CM CHRISTMAS GLASS BALL 20 LIGHTS        12   \n",
       "1  489434    79323P                   PINK CHERRY LIGHTS        12   \n",
       "2  489434    79323W                  WHITE CHERRY LIGHTS        12   \n",
       "3  489434     22041         RECORD FRAME 7\" SINGLE SIZE         48   \n",
       "4  489434     21232       STRAWBERRY CERAMIC TRINKET BOX        24   \n",
       "\n",
       "          InvoiceDate  Price  Customer ID         Country  Amount  Holiday  \\\n",
       "0 2009-12-01 07:45:00   6.95      13085.0  United Kingdom    83.4    False   \n",
       "1 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0    False   \n",
       "2 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0    False   \n",
       "3 2009-12-01 07:45:00   2.10      13085.0  United Kingdom   100.8    False   \n",
       "4 2009-12-01 07:45:00   1.25      13085.0  United Kingdom    30.0    False   \n",
       "\n",
       "  HolidayName  \n",
       "0        None  \n",
       "1        None  \n",
       "2        None  \n",
       "3        None  \n",
       "4        None  "
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this code snippet, we use the holidays package to create several holiday-related features:\n",
    "\n",
    "**Holiday:** We create a binary feature indicating if a transaction occurred on a holiday in the United State. The holidays.UK() function generates a holiday object for the UK, and we check if each transaction date is present in the holiday object.\n",
    "\n",
    "**HolidayName:** We create a feature indicating the name of the holiday for each transaction (if applicable). If a transaction occurred on a holiday, the feature will contain the name of the holiday; otherwise, it will be None.\n",
    "\n",
    "By leveraging the holidays package, we can easily generate holiday-related features for our dataset without manually defining the holiday dates. The package supports various countries and regions, so we can choose the appropriate holiday calendar based on our data and analysis requirements.  \n",
    "\n",
    "These holiday-related features can provide valuable insights into customer purchasing behavior during holidays and weekends, allowing us to create more targeted marketing strategies and campaigns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.4 Product-based Features\n",
    "Product-based features capture information about the products customers purchase and their characteristics. Let's create some product-based features for our Online Retail dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Holiday</th>\n",
       "      <th>HolidayName</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>489434</td>\n",
       "      <td>85048</td>\n",
       "      <td>15CM CHRISTMAS GLASS BALL 20 LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.95</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>83.4</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323P</td>\n",
       "      <td>PINK CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>489434</td>\n",
       "      <td>79323W</td>\n",
       "      <td>WHITE CHERRY LIGHTS</td>\n",
       "      <td>12</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>6.75</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>81.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>489434</td>\n",
       "      <td>22041</td>\n",
       "      <td>RECORD FRAME 7\" SINGLE SIZE</td>\n",
       "      <td>48</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>2.10</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>100.8</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>489434</td>\n",
       "      <td>21232</td>\n",
       "      <td>STRAWBERRY CERAMIC TRINKET BOX</td>\n",
       "      <td>24</td>\n",
       "      <td>2009-12-01 07:45:00</td>\n",
       "      <td>1.25</td>\n",
       "      <td>13085.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "      <td>30.0</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Invoice StockCode                          Description  Quantity  \\\n",
       "0  489434     85048  15CM CHRISTMAS GLASS BALL 20 LIGHTS        12   \n",
       "1  489434    79323P                   PINK CHERRY LIGHTS        12   \n",
       "2  489434    79323W                  WHITE CHERRY LIGHTS        12   \n",
       "3  489434     22041         RECORD FRAME 7\" SINGLE SIZE         48   \n",
       "4  489434     21232       STRAWBERRY CERAMIC TRINKET BOX        24   \n",
       "\n",
       "          InvoiceDate  Price  Customer ID         Country  Amount  Holiday  \\\n",
       "0 2009-12-01 07:45:00   6.95      13085.0  United Kingdom    83.4    False   \n",
       "1 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0    False   \n",
       "2 2009-12-01 07:45:00   6.75      13085.0  United Kingdom    81.0    False   \n",
       "3 2009-12-01 07:45:00   2.10      13085.0  United Kingdom   100.8    False   \n",
       "4 2009-12-01 07:45:00   1.25      13085.0  United Kingdom    30.0    False   \n",
       "\n",
       "  HolidayName  \n",
       "0        None  \n",
       "1        None  \n",
       "2        None  \n",
       "3        None  \n",
       "4        None  "
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "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>NumUniqueProducts</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>TotalQuantity</th>\n",
       "      <th>NumInvoices</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12346.0</th>\n",
       "      <td>30</td>\n",
       "      <td>-64.68</td>\n",
       "      <td>52</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>126</td>\n",
       "      <td>5633.32</td>\n",
       "      <td>3286</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>25</td>\n",
       "      <td>2019.40</td>\n",
       "      <td>2714</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12349.0</th>\n",
       "      <td>139</td>\n",
       "      <td>4404.54</td>\n",
       "      <td>1619</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12350.0</th>\n",
       "      <td>17</td>\n",
       "      <td>334.40</td>\n",
       "      <td>197</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             NumUniqueProducts  TotalSales  TotalQuantity  NumInvoices\n",
       "Customer ID                                                           \n",
       "12346.0                     30      -64.68             52           17\n",
       "12347.0                    126     5633.32           3286            8\n",
       "12348.0                     25     2019.40           2714            5\n",
       "12349.0                    139     4404.54           1619            5\n",
       "12350.0                     17      334.40            197            1"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pbf = df.groupby('Customer ID')[['StockCode', 'Amount', 'Quantity', 'Invoice']].agg({'StockCode': 'nunique', 'Amount': 'sum', 'Quantity': 'sum', 'Invoice': 'nunique'})\n",
    "                                                                 \n",
    "# rename the column StockCode to 'NumUniqueProduts'\n",
    "df_pbf = df_pbf.rename(columns={'StockCode': 'NumUniqueProducts', 'Amount': 'TotalSales', 'Quantity': 'TotalQuantity', 'Invoice': 'NumInvoices'})\n",
    "df_pbf.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this code snippet, we calculate various product-based features such as the total number of unique products purchased by each customer, the average unit price of products purchased, the total quantity of products purchased, and the average quantity per transaction.\n",
    "\n",
    "These features can help us understand customer preferences and purchasing habits, which can inform product recommendations, cross-selling, and upselling strategies.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.5 Interaction Features \n",
    "\n",
    "Interaction features capture the relationship between two or more variables in our dataset. For example, we might want to understand how the relationship between a customer's total sales and the number of orders they place impacts their overall value to the business.\n",
    "\n",
    "To create interaction features for our Online Retail dataset, we can use techniques such as:\n",
    "\n",
    "* Multiplying or dividing features to create new features that capture the relationship between them\n",
    "* Polynomial features to capture more complex relationships between variables\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.1 Interaction features by division and multiplication"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Let's create interaction features for our Online Retail dataset:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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>NumUniqueProducts</th>\n",
       "      <th>TotalSales</th>\n",
       "      <th>TotalQuantity</th>\n",
       "      <th>NumInvoices</th>\n",
       "      <th>AvgUnitPrice</th>\n",
       "      <th>AvgQuantityPerTransaction</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12346.0</th>\n",
       "      <td>30</td>\n",
       "      <td>-64.68</td>\n",
       "      <td>52</td>\n",
       "      <td>17</td>\n",
       "      <td>-2.156000</td>\n",
       "      <td>3.058824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>126</td>\n",
       "      <td>5633.32</td>\n",
       "      <td>3286</td>\n",
       "      <td>8</td>\n",
       "      <td>44.708889</td>\n",
       "      <td>410.750000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>25</td>\n",
       "      <td>2019.40</td>\n",
       "      <td>2714</td>\n",
       "      <td>5</td>\n",
       "      <td>80.776000</td>\n",
       "      <td>542.800000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12349.0</th>\n",
       "      <td>139</td>\n",
       "      <td>4404.54</td>\n",
       "      <td>1619</td>\n",
       "      <td>5</td>\n",
       "      <td>31.687338</td>\n",
       "      <td>323.800000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12350.0</th>\n",
       "      <td>17</td>\n",
       "      <td>334.40</td>\n",
       "      <td>197</td>\n",
       "      <td>1</td>\n",
       "      <td>19.670588</td>\n",
       "      <td>197.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             NumUniqueProducts  TotalSales  TotalQuantity  NumInvoices  \\\n",
       "Customer ID                                                              \n",
       "12346.0                     30      -64.68             52           17   \n",
       "12347.0                    126     5633.32           3286            8   \n",
       "12348.0                     25     2019.40           2714            5   \n",
       "12349.0                    139     4404.54           1619            5   \n",
       "12350.0                     17      334.40            197            1   \n",
       "\n",
       "             AvgUnitPrice  AvgQuantityPerTransaction  \n",
       "Customer ID                                           \n",
       "12346.0         -2.156000                   3.058824  \n",
       "12347.0         44.708889                 410.750000  \n",
       "12348.0         80.776000                 542.800000  \n",
       "12349.0         31.687338                 323.800000  \n",
       "12350.0         19.670588                 197.000000  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate the average unit price of products purchased by each customer\n",
    "df_pbf['AvgUnitPrice'] = df_pbf['TotalSales'] /df_pbf['NumUniqueProducts']\n",
    "\n",
    "# Calculate the average quantity per transaction for each customer\n",
    "df_pbf['AvgQuantityPerTransaction'] = df_pbf['TotalQuantity'] /df_pbf['NumInvoices']\n",
    "\n",
    "df_pbf.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code snippet creates an interaction feature called 'AvgUnitPrice' by combining two existing features: 'TotalSales' and 'NumUniqueProducts'. The interaction is achieved through the division operation.\n",
    "\n",
    "The 'TotalSales' feature represents the total revenue generated by each customer, while 'NumUniqueProducts' represents the number of unique products purchased by each customer. By dividing 'TotalSales' by 'NumUniqueProducts', we create a new feature that captures the relationship between these two variables.\n",
    "\n",
    "The resulting 'AvgUnitPrice' feature provides insights into the average spending per product for each customer. It helps understand how the combination of total sales and the number of unique products impacts the average unit price, which can be an important factor in analyzing customer behavior and value.\n",
    "\n",
    "The interaction feature 'AvgQuantityPerTransaction' provides insights into the average number of products a customer buys in each transaction. It helps understand how the combination of total quantity and the number of transactions impacts the average quantity per transaction, which can be useful in analyzing customer purchasing patterns and behavior."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.2 Polynomial features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generating polynomial features involves creating new features by combining existing features with polynomial functions, such as squares, cubes, or interactions . Scikit-learn's PolynomialFeatures class is a tool for generating these features . Feature engineering, including the generation of polynomial features, is crucial for improving model performance and can help avoid overfitting and underfitting problems."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "\n",
    "# Assume you have a DataFrame called 'df' with the following columns:\n",
    "# 'SalesVolume', 'DiscountPercentage', 'Recency', 'Frequency', 'MarketingSpend', 'ConversionRate'\n",
    "\n",
    "# Select the relevant features for polynomial feature creation\n",
    "features = ['Recency', 'Frequency']\n",
    "X = rfm[features]\n",
    "\n",
    "# Create polynomial features of degree 2\n",
    "poly = PolynomialFeatures(degree=2, include_bias=False)\n",
    "X_poly = poly.fit_transform(X)\n",
    "\n",
    "# Get the feature names\n",
    "feature_names = poly.get_feature_names_out(features)\n",
    "\n",
    "# Create a new DataFrame with the polynomial features\n",
    "df_poly = pd.DataFrame(X_poly, columns=feature_names)\n",
    "\n",
    "# Concatenate the polynomial features with the original DataFrame\n",
    "rfm_enhanced = pd.concat([rfm.reset_index(), df_poly.reset_index()], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "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>Customer ID</th>\n",
       "      <th>Recency</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "      <th>R_rank</th>\n",
       "      <th>F_rank</th>\n",
       "      <th>M_rank</th>\n",
       "      <th>RFM_score</th>\n",
       "      <th>index</th>\n",
       "      <th>Recency</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Recency^2</th>\n",
       "      <th>Recency Frequency</th>\n",
       "      <th>Frequency^2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12346.0</td>\n",
       "      <td>325</td>\n",
       "      <td>17</td>\n",
       "      <td>-64.68</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>441</td>\n",
       "      <td>0</td>\n",
       "      <td>325.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>105625.0</td>\n",
       "      <td>5525.0</td>\n",
       "      <td>289.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12347.0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>5633.32</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>135</td>\n",
       "      <td>1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>64.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12348.0</td>\n",
       "      <td>75</td>\n",
       "      <td>5</td>\n",
       "      <td>2019.40</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>324</td>\n",
       "      <td>2</td>\n",
       "      <td>75.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5625.0</td>\n",
       "      <td>375.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>12349.0</td>\n",
       "      <td>18</td>\n",
       "      <td>5</td>\n",
       "      <td>4404.54</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>125</td>\n",
       "      <td>3</td>\n",
       "      <td>18.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>324.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12350.0</td>\n",
       "      <td>310</td>\n",
       "      <td>1</td>\n",
       "      <td>334.40</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>412</td>\n",
       "      <td>4</td>\n",
       "      <td>310.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>96100.0</td>\n",
       "      <td>310.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Customer ID  Recency  Frequency  Monetary  R_rank  F_rank  M_rank  \\\n",
       "0      12346.0      325         17    -64.68       4       4       1   \n",
       "1      12347.0        2          8   5633.32       1       3       5   \n",
       "2      12348.0       75          5   2019.40       3       2       4   \n",
       "3      12349.0       18          5   4404.54       1       2       5   \n",
       "4      12350.0      310          1    334.40       4       1       2   \n",
       "\n",
       "  RFM_score  index  Recency  Frequency  Recency^2  Recency Frequency  \\\n",
       "0       441      0    325.0       17.0   105625.0             5525.0   \n",
       "1       135      1      2.0        8.0        4.0               16.0   \n",
       "2       324      2     75.0        5.0     5625.0              375.0   \n",
       "3       125      3     18.0        5.0      324.0               90.0   \n",
       "4       412      4    310.0        1.0    96100.0              310.0   \n",
       "\n",
       "   Frequency^2  \n",
       "0        289.0  \n",
       "1         64.0  \n",
       "2         25.0  \n",
       "3         25.0  \n",
       "4          1.0  "
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfm_enhanced.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The code snippet generates polynomial features: 'Recency^2', 'Recency*Frequency', and 'Frequency^2'. These capture the interactions and higher-order relationships between the original 'Recency' and 'Frequency' features. Incorporating polynomial features can offer additional insights for marketing and sales analysis, potentially enhancing predictive model performance. However, it's important to consider the interpretability and relevance of these polynomial features within the specific context of our marketing and sales domain. We must keep in mind that the ultimate goal of feature engineering is to provide actionable insights for decision-makers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.6 Conclusion\n",
    "In this chapter, we explored feature engineering techniques for marketing data using Python. We created new features based on RFM analysis, time-based patterns, product characteristics, and interaction features to capture more complex relationships in our data.\n",
    "\n",
    "Feature engineering is an iterative process that requires domain knowledge, creativity, and experimentation. By creating informative and relevant features, we can improve the performance of our marketing analytics models and gain deeper actionable insights into customer behavior.\n",
    "\n",
    "In the next chapter, we'll dive into predictive analytics and learn how to build machine learning models for marketing tasks such as customer segmentation, churn prediction, and sales forecasting."
   ]
  }
 ],
 "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": "3.7.16"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}