20. Ads Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "42bc715e",
   "metadata": {},
   "source": [
    "# Goal of Analysis\n",
    "\n",
    "## 1. If you had to identify the 5 best ad groups, which ones would they be? Which metric did you choose to identify the best ad groups? Why? Explain the pros and cons of your metrics.\n",
    "\n",
    "## 2. For each group, predict how many ads will be shown on Dec, 15 (assume each ad group keeps following its trend).\n",
    "\n",
    "## 3. Cluster ads into 3 groups: the ones whose avg_cost_per_click is going up, the ones whose avg_cost_per_click is flat and the ones whose avg_cost_per_click is going down."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "dcfe5fc0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7775be4b",
   "metadata": {},
   "source": [
    "### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "92cf9699",
   "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>date</th>\n",
       "      <th>shown</th>\n",
       "      <th>clicked</th>\n",
       "      <th>converted</th>\n",
       "      <th>avg_cost_per_click</th>\n",
       "      <th>total_revenue</th>\n",
       "      <th>ad</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-10-01</td>\n",
       "      <td>65877</td>\n",
       "      <td>2339</td>\n",
       "      <td>43</td>\n",
       "      <td>0.90</td>\n",
       "      <td>641.62</td>\n",
       "      <td>ad_group_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-10-02</td>\n",
       "      <td>65100</td>\n",
       "      <td>2498</td>\n",
       "      <td>38</td>\n",
       "      <td>0.94</td>\n",
       "      <td>756.37</td>\n",
       "      <td>ad_group_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>70658</td>\n",
       "      <td>2313</td>\n",
       "      <td>49</td>\n",
       "      <td>0.86</td>\n",
       "      <td>970.90</td>\n",
       "      <td>ad_group_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-10-04</td>\n",
       "      <td>69809</td>\n",
       "      <td>2833</td>\n",
       "      <td>51</td>\n",
       "      <td>1.01</td>\n",
       "      <td>907.39</td>\n",
       "      <td>ad_group_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-10-05</td>\n",
       "      <td>68186</td>\n",
       "      <td>2696</td>\n",
       "      <td>41</td>\n",
       "      <td>1.00</td>\n",
       "      <td>879.45</td>\n",
       "      <td>ad_group_1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date  shown  clicked  converted  avg_cost_per_click  total_revenue  \\\n",
       "0 2015-10-01  65877     2339         43                0.90         641.62   \n",
       "1 2015-10-02  65100     2498         38                0.94         756.37   \n",
       "2 2015-10-03  70658     2313         49                0.86         970.90   \n",
       "3 2015-10-04  69809     2833         51                1.01         907.39   \n",
       "4 2015-10-05  68186     2696         41                1.00         879.45   \n",
       "\n",
       "           ad  \n",
       "0  ad_group_1  \n",
       "1  ad_group_1  \n",
       "2  ad_group_1  \n",
       "3  ad_group_1  \n",
       "4  ad_group_1  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('ad_table.csv', parse_dates = ['date'])\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "5120e703",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 2115 entries, 0 to 2114\n",
      "Data columns (total 7 columns):\n",
      " #   Column              Non-Null Count  Dtype         \n",
      "---  ------              --------------  -----         \n",
      " 0   date                2115 non-null   datetime64[ns]\n",
      " 1   shown               2115 non-null   int64         \n",
      " 2   clicked             2115 non-null   int64         \n",
      " 3   converted           2115 non-null   int64         \n",
      " 4   avg_cost_per_click  2115 non-null   float64       \n",
      " 5   total_revenue       2115 non-null   float64       \n",
      " 6   ad                  2115 non-null   object        \n",
      "dtypes: datetime64[ns](1), float64(2), int64(3), object(1)\n",
      "memory usage: 115.8+ KB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "839cdbb7",
   "metadata": {},
   "source": [
    "### Question 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "31fc3b1f",
   "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>date</th>\n",
       "      <th>shown</th>\n",
       "      <th>clicked</th>\n",
       "      <th>converted</th>\n",
       "      <th>avg_cost_per_click</th>\n",
       "      <th>total_revenue</th>\n",
       "      <th>ad</th>\n",
       "      <th>profit</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-10-01</td>\n",
       "      <td>65877</td>\n",
       "      <td>2339</td>\n",
       "      <td>43</td>\n",
       "      <td>0.90</td>\n",
       "      <td>641.62</td>\n",
       "      <td>ad_group_1</td>\n",
       "      <td>-1463.48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-10-02</td>\n",
       "      <td>65100</td>\n",
       "      <td>2498</td>\n",
       "      <td>38</td>\n",
       "      <td>0.94</td>\n",
       "      <td>756.37</td>\n",
       "      <td>ad_group_1</td>\n",
       "      <td>-1591.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>70658</td>\n",
       "      <td>2313</td>\n",
       "      <td>49</td>\n",
       "      <td>0.86</td>\n",
       "      <td>970.90</td>\n",
       "      <td>ad_group_1</td>\n",
       "      <td>-1018.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-10-04</td>\n",
       "      <td>69809</td>\n",
       "      <td>2833</td>\n",
       "      <td>51</td>\n",
       "      <td>1.01</td>\n",
       "      <td>907.39</td>\n",
       "      <td>ad_group_1</td>\n",
       "      <td>-1953.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-10-05</td>\n",
       "      <td>68186</td>\n",
       "      <td>2696</td>\n",
       "      <td>41</td>\n",
       "      <td>1.00</td>\n",
       "      <td>879.45</td>\n",
       "      <td>ad_group_1</td>\n",
       "      <td>-1816.55</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date  shown  clicked  converted  avg_cost_per_click  total_revenue  \\\n",
       "0 2015-10-01  65877     2339         43                0.90         641.62   \n",
       "1 2015-10-02  65100     2498         38                0.94         756.37   \n",
       "2 2015-10-03  70658     2313         49                0.86         970.90   \n",
       "3 2015-10-04  69809     2833         51                1.01         907.39   \n",
       "4 2015-10-05  68186     2696         41                1.00         879.45   \n",
       "\n",
       "           ad   profit  \n",
       "0  ad_group_1 -1463.48  \n",
       "1  ad_group_1 -1591.75  \n",
       "2  ad_group_1 -1018.28  \n",
       "3  ad_group_1 -1953.94  \n",
       "4  ad_group_1 -1816.55  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate the profit: total_revenue - total_cost\n",
    "\n",
    "data['profit'] = data['total_revenue'] - data['avg_cost_per_click'] * data['clicked']\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c09b0174",
   "metadata": {},
   "source": [
    "Because I already calculated the net profit using the total revenue and the total cost, so I will use profit as the metric to identify the 5 best ad groups."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6102e8b",
   "metadata": {},
   "source": [
    "I first difine a function to calculate the unit profit using the proft and shown."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "4c7f90b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "def unit_profit(df):\n",
    "    profit = df['profit'].sum()\n",
    "    shown = df['shown'].sum()\n",
    "    val = profit / shown\n",
    "    \n",
    "    return val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c9061fb9",
   "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>ad</th>\n",
       "      <th>unit_profit</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ad_group_16</td>\n",
       "      <td>0.020516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>ad_group_2</td>\n",
       "      <td>0.020116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ad_group_14</td>\n",
       "      <td>0.016485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>ad_group_31</td>\n",
       "      <td>0.005824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>ad_group_27</td>\n",
       "      <td>0.000725</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             ad  unit_profit\n",
       "7   ad_group_16     0.020516\n",
       "11   ad_group_2     0.020116\n",
       "5   ad_group_14     0.016485\n",
       "24  ad_group_31     0.005824\n",
       "19  ad_group_27     0.000725"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grouped = data.groupby('ad').apply(unit_profit).reset_index()\n",
    "grouped = grouped.rename(columns={0: 'unit_profit'})\n",
    "grouped = grouped.sort_values(by='unit_profit', ascending=False)\n",
    "\n",
    "grouped.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "969db477",
   "metadata": {},
   "source": [
    "Based on above calculation, we can see the the top 5 best ad groups are: ad_group_16, ad_group_2, ad_group_14, ad_group_31, ad_group_27."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc227c4d",
   "metadata": {},
   "source": [
    "### Question 3"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe188be3",
   "metadata": {},
   "source": [
    "I need to define a function to calculate the avg_cost_per_click trend before answering the question"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "edc8d055",
   "metadata": {},
   "outputs": [],
   "source": [
    "def cost_stats(df):\n",
    "    tmp = df.sort_values(by='date')['avg_cost_per_click'].values\n",
    "    ratio = tmp[1:] / tmp[:-1]\n",
    "    \n",
    "    ratio_mean = np.mean(ratio)\n",
    "    ratio_min = np.min(ratio)\n",
    "    ratio_25 = np.percentile(ratio, 25)\n",
    "    ratio_50 = np.percentile(ratio, 50)\n",
    "    ratio_75 = np.percentile(ratio, 75)\n",
    "    ratio_max = np.max(ratio)\n",
    "    \n",
    "    return pd.Series([ratio_mean, ratio_min, ratio_25, ratio_50, ratio_75, ratio_max], \n",
    "                     index=['mean', 'min', '25%', '50%', '75%', 'max'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "26ad8081",
   "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>mean</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad</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>ad_group_1</th>\n",
       "      <td>1.007436</td>\n",
       "      <td>0.822430</td>\n",
       "      <td>0.930747</td>\n",
       "      <td>0.989999</td>\n",
       "      <td>1.094223</td>\n",
       "      <td>1.282353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_group_10</th>\n",
       "      <td>1.011576</td>\n",
       "      <td>0.689922</td>\n",
       "      <td>0.920666</td>\n",
       "      <td>0.976923</td>\n",
       "      <td>1.105323</td>\n",
       "      <td>1.382022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_group_11</th>\n",
       "      <td>inf</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.957944</td>\n",
       "      <td>1.002941</td>\n",
       "      <td>1.052941</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_group_12</th>\n",
       "      <td>inf</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.945670</td>\n",
       "      <td>1.005051</td>\n",
       "      <td>1.060455</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_group_13</th>\n",
       "      <td>inf</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.958716</td>\n",
       "      <td>1.000015</td>\n",
       "      <td>1.071825</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 mean       min       25%       50%       75%       max\n",
       "ad                                                                     \n",
       "ad_group_1   1.007436  0.822430  0.930747  0.989999  1.094223  1.282353\n",
       "ad_group_10  1.011576  0.689922  0.920666  0.976923  1.105323  1.382022\n",
       "ad_group_11       inf  0.000000  0.957944  1.002941  1.052941       inf\n",
       "ad_group_12       inf  0.000000  0.945670  1.005051  1.060455       inf\n",
       "ad_group_13       inf  0.000000  0.958716  1.000015  1.071825       inf"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats = data.groupby('ad').apply(cost_stats)\n",
    "stats.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "f6fc0447",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualization\n",
    "\n",
    "hist_kws={'histtype': 'bar', 'edgecolor':'black', 'alpha': 0.2}\n",
    "\n",
    "fig, ax = plt.subplots(nrows=3, ncols=1, figsize=(12, 10), sharex=True)\n",
    "sns.distplot(stats['25%'], bins=40, ax=ax[0], label='25%', hist_kws=hist_kws)\n",
    "ax[0].legend(fontsize=12)\n",
    "sns.distplot(stats['50%'], bins=40, ax=ax[1], label='50%', hist_kws=hist_kws)\n",
    "ax[1].legend(fontsize=12)\n",
    "sns.distplot(stats['75%'], bins=40, ax=ax[2], label='75%', hist_kws=hist_kws)\n",
    "ax[2].legend(fontsize=12)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23b584d3",
   "metadata": {},
   "source": [
    "Based on the above analysis, we should manually set a series cutoff thresholds to identify the cluster of going up, flat, and going down."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0736e131",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}