success-contacts.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "###\n",
    "# This code answer the questions:\n",
    "# Making a relationship between the number of contacts and campaign success\n",
    "# Based on the results of this campaign, what are the average and the maximum number of responses \n",
    "# you indicate to optimize adherence? what are the relevant points to be observed?\n",
    "# By Sidon 2018\n",
    "###"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Null Hypothesis:<br>\n",
    "There is no relationship between the number of contacts and campaign success.<br>\n",
    "H0 = There is no relationship between the two categorical variables - they are indepent.\n",
    "\n",
    "Alternative Hypothesis:<br>\n",
    "There is relationship between the number of contacts and campaign success.<br>\n",
    "H1 = There is a relationship between the two categorical variables - they are not idependent.\n",
    "\n",
    "H0 = proportion of number of contacts = proportion of success<br>\n",
    "h1 = proportion of number of contacts != proportion of success\n",
    "\n",
    "chi-square test of independence:<br>\n",
    "The idea behind the chi-square test of independence, much like the analysis of variance is to measure how far the data are from what is claimed in the null hypothesis. The further the data are from the null hypothesis, the more evidence the data presents against it. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "from scipy import stats\n",
    "from tabulate import tabulate, tabulate_formats\n",
    "from utils import Dataset\n",
    "from collections import OrderedDict\n",
    "%matplotlib inline "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "ds =  Dataset()\n",
    "data = ds.get_data(['campaign','previous','y'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Inserir coluna para totalizar os contatos realizados\n",
    "tt_contacts = data.campaign+data['previous'] \n",
    "data.insert(loc=2, column='tt_contacts', value=tt_contacts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Copy for conveniency\n",
    "data1 = data.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remove outliers\n",
    "q = data1['tt_contacts'].quantile(0.99)\n",
    "data1 = data1[data1['tt_contacts']<q]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>campaign</th>\n",
       "      <th>previous</th>\n",
       "      <th>tt_contacts</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45206</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45207</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45208</th>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45209</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45210</th>\n",
       "      <td>2</td>\n",
       "      <td>11</td>\n",
       "      <td>13</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>44683 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       campaign  previous  tt_contacts    y\n",
       "0             1         0            1   no\n",
       "1             1         0            1   no\n",
       "2             1         0            1   no\n",
       "3             1         0            1   no\n",
       "4             1         0            1   no\n",
       "...         ...       ...          ...  ...\n",
       "45206         3         0            3  yes\n",
       "45207         2         0            2  yes\n",
       "45208         5         3            8  yes\n",
       "45209         4         0            4   no\n",
       "45210         2        11           13   no\n",
       "\n",
       "[44683 rows x 4 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# contingency table of observed counts\n",
    "ct_total = pd.crosstab(data1.y, data1.tt_contacts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "``` \n",
    "lambda_ : float or str, optional\n",
    "\n",
    "lambda_ gives the power in the Cressie-Read power divergence statistic. The default is 1. For convenience, lambda_ may be assigned one of the following strings, in which case the corresponding numerical value is used:\n",
    "\n",
    "String              Value   Description\n",
    "\"pearson\"             1     Pearson's chi-squared statistic.\n",
    "                            In this case, the function is\n",
    "                            equivalent to `stats.chisquare`.\n",
    "\"log-likelihood\"      0     Log-likelihood ratio. Also known as\n",
    "                            the G-test [R256]_.\n",
    "\"freeman-tukey\"      -1/2   Freeman-Tukey statistic.\n",
    "\"mod-log-likelihood\" -1     Modified log-likelihood ratio.\n",
    "\"neyman\"             -2     Neyman's statistic.\n",
    "\"cressie-read\"        2/3   The power recommended in [R258]_.\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# G-test for totals\n",
    "chi2g = {}\n",
    "cs1 = stats.chi2_contingency(ct_total, lambda_=\"log-likelihood\")\n",
    "chi2g['chi-square'] = [cs1[0]]\n",
    "chi2g['p-value'] = [cs1[1]]\n",
    "chi2g['dof'] = [cs1[2]]\n",
    "# chi2g['expected'] = [cs1[3]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log-likelihood ratio. Also known as the G-test. for totals contacts x success\n",
      "╒══════════════╤═════════════╤═══════╕\n",
      "│  chi-square  │   p-value   │  dof  │\n",
      "╞══════════════╪═════════════╪═══════╡\n",
      "│   60.1768    │ 4.88676e-07 │  16   │\n",
      "╘══════════════╧═════════════╧═══════╛\n"
     ]
    }
   ],
   "source": [
    "print ('Log-likelihood ratio. Also known as the G-test. for totals contacts x success')\n",
    "print (tabulate(chi2g, tablefmt='fancy_grid', headers='keys', numalign=\"center\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7feaf0beeef0>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ct_total.T.plot(figsize=(9, 7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# column percentages\n",
    "colsum=ct_total.sum(axis=0)\n",
    "colsum.astype(int)\n",
    "colpct=ct_total/colsum\n",
    "\n",
    "colpct_t = colpct.T.sort_values(by=['yes'], ascending=False)\n",
    "colpct_t.insert(loc=0, column='Freq.', value=colsum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Totals contacts frequency (campaing + previous)\n",
      "╒═══════════════╤═════════╤══════════╤═══════════╕\n",
      "│   tt_contacts │   Freq. │       no │       yes │\n",
      "╞═══════════════╪═════════╪══════════╪═══════════╡\n",
      "│             9 │     546 │ 0.842491 │ 0.157509  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            11 │     321 │ 0.853583 │ 0.146417  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             5 │    2514 │ 0.863564 │ 0.136436  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             6 │    1819 │ 0.86641  │ 0.13359   │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             7 │    1099 │ 0.867152 │ 0.132848  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            10 │     446 │ 0.869955 │ 0.130045  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             4 │    4419 │ 0.871238 │ 0.128762  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             3 │    6449 │ 0.875174 │ 0.124826  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             8 │     834 │ 0.8753   │ 0.1247    │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            12 │     247 │ 0.88664  │ 0.11336   │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             1 │   13529 │ 0.887427 │ 0.112573  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            15 │     134 │ 0.88806  │ 0.11194   │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            14 │     145 │ 0.889655 │ 0.110345  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│             2 │   11760 │ 0.893963 │ 0.106037  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            13 │     209 │ 0.894737 │ 0.105263  │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            16 │     115 │ 0.930435 │ 0.0695652 │\n",
      "├───────────────┼─────────┼──────────┼───────────┤\n",
      "│            17 │      97 │ 0.938144 │ 0.0618557 │\n",
      "╘═══════════════╧═════════╧══════════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "print('Totals contacts frequency (campaing + previous)')\n",
    "print(tabulate(colpct_t, tablefmt='fancy_grid', headers='keys', numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Freeman-Turkey test for totals\n",
    "chi2ft = {}\n",
    "cs2 = stats.chi2_contingency(ct_total, lambda_=\"freeman-tukey\")\n",
    "chi2ft['chi-square'] = [cs2[0]]\n",
    "chi2ft['p-value'] = [cs2[1]]\n",
    "chi2ft['dof'] = [cs2[2]]\n",
    "# chi2ft['expected'] = [cs2[3]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Freeman-Tukey statistic. for totals contacts x success\n",
      "╒══════════════╤═════════════╤═══════╕\n",
      "│  chi-square  │   p-value   │  dof  │\n",
      "╞══════════════╪═════════════╪═══════╡\n",
      "│    60.278    │ 4.69836e-07 │  16   │\n",
      "╘══════════════╧═════════════╧═══════╛\n"
     ]
    }
   ],
   "source": [
    "print('Freeman-Tukey statistic. for totals contacts x success')\n",
    "print(tabulate(chi2ft, tablefmt='fancy_grid', headers='keys', numalign=\"center\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "╒════╤══════╤════════════╤════════════╤═══════╤═══════╤═════════════╤═══════════╕\n",
      "│    │   tt │   campaign │   previous │    no │   yes │   previous% │   success │\n",
      "╞════╪══════╪════════════╪════════════╪═══════╪═══════╪═════════════╪═══════════╡\n",
      "│  8 │    9 │         90 │         72 │   460 │    86 │    0.444444 │ 0.157509  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 10 │   11 │        105 │        104 │   274 │    47 │    0.497608 │ 0.146417  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  4 │    5 │         30 │         20 │  2171 │   343 │    0.4      │ 0.136436  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  5 │    6 │         42 │         30 │  1576 │   243 │    0.416667 │ 0.13359   │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  6 │    7 │         56 │         42 │   953 │   146 │    0.428571 │ 0.132848  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  9 │   10 │        101 │         89 │   388 │    58 │    0.468421 │ 0.130045  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  3 │    4 │         20 │         12 │  3850 │   569 │    0.375    │ 0.128762  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  2 │    3 │         12 │          6 │  5644 │   805 │    0.333333 │ 0.124826  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  7 │    8 │         72 │         56 │   730 │   104 │    0.4375   │ 0.1247    │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 11 │   12 │        109 │        119 │   219 │    28 │    0.52193  │ 0.11336   │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  0 │    1 │          2 │          0 │ 12006 │  1523 │    0        │ 0.112573  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 14 │   15 │        114 │        156 │   119 │    15 │    0.577778 │ 0.11194   │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 13 │   14 │        134 │        146 │   129 │    16 │    0.521429 │ 0.110345  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│  1 │    2 │          6 │          2 │ 10513 │  1247 │    0.25     │ 0.106037  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 12 │   13 │        124 │        123 │   187 │    22 │    0.497976 │ 0.105263  │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 15 │   16 │        106 │        134 │   107 │     8 │    0.558333 │ 0.0695652 │\n",
      "├────┼──────┼────────────┼────────────┼───────┼───────┼─────────────┼───────────┤\n",
      "│ 16 │   17 │         91 │        113 │    91 │     6 │    0.553922 │ 0.0618557 │\n",
      "╘════╧══════╧════════════╧════════════╧═══════╧═══════╧═════════════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "ct = pd.DataFrame(data1.groupby(by=['tt_contacts','campaign','previous', 'y']).size())\n",
    "def totaldf():\n",
    "    d1 = OrderedDict()\n",
    "    for c in ct.iterrows():\n",
    "        tt = c[0][0]\n",
    "        response = c[0][3]\n",
    "        if tt in d1:\n",
    "            if response in d1[tt]:\n",
    "                d1[tt][response]+=c[1][0]\n",
    "            else:\n",
    "                d1[tt][response]=c[1][0]\n",
    "            d1[tt]['campaign']+=c[0][1]    \n",
    "            d1[tt]['previous']+=c[0][2]    \n",
    "        else:\n",
    "            d1[tt]={'tt': c[0][0], 'campaign': c[0][1], 'previous': c[0][2], response: c[1][0]}\n",
    "        lst = [d1[d] for d in d1]\n",
    "        df = pd.DataFrame(lst, columns=['tt', 'campaign', 'previous', 'no', 'yes'])\n",
    "    df['previous%'] = df.previous/(df.campaign+df.previous)\n",
    "    df['success'] = df.yes/(df.yes+df.no)\n",
    "    return df\n",
    "\n",
    "df = totaldf().sort_values(by=['success'], ascending=False)    \n",
    "print(tabulate(df, tablefmt='fancy_grid', headers='keys', numalign=\"decimal\"))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rank considering only previous' calls\n",
      "╒════════════╤═══════╤═══════╤═══════════╕\n",
      "│   previous │    no │   yes │   success │\n",
      "╞════════════╪═══════╪═══════╪═══════════╡\n",
      "│         10 │    40 │    26 │ 0.393939  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         13 │    18 │     9 │ 0.333333  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         14 │     9 │     4 │ 0.307692  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          8 │    90 │    39 │ 0.302326  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          6 │   193 │    83 │ 0.300725  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          5 │   338 │   121 │ 0.263617  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          7 │   151 │    54 │ 0.263415  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          9 │    68 │    24 │ 0.26087   │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          3 │   848 │   294 │ 0.257443  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          4 │   543 │   171 │ 0.239496  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         11 │    46 │    14 │ 0.233333  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          2 │  1650 │   456 │ 0.216524  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          1 │  2189 │   583 │ 0.210317  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         12 │    31 │     8 │ 0.205128  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         15 │     7 │     1 │ 0.125     │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          0 │ 33193 │  3379 │ 0.0923931 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         16 │     3 │     0 │ 0         │\n",
      "╘════════════╧═══════╧═══════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "dfp = pd.crosstab(data1.y, [data1.previous]).T\n",
    "dfp['success']=dfp.yes/(dfp.no+dfp.yes) \n",
    "dfp = dfp.sort_values(by=['success'], ascending=False)\n",
    "print(\"Rank considering only previous' calls\")\n",
    "print(tabulate(dfp, tablefmt='fancy_grid', headers='keys', numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rank considering only campaign's calls\n",
      "╒════════════╤═══════╤═══════╤═══════════╕\n",
      "│   campaign │    no │   yes │   success │\n",
      "╞════════════╪═══════╪═══════╪═══════════╡\n",
      "│          1 │ 14962 │  2559 │ 0.146053  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          2 │ 11077 │  1395 │ 0.111851  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          3 │  4886 │   614 │ 0.111636  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          4 │  3195 │   314 │ 0.0894842 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         17 │    63 │     6 │ 0.0869565 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         11 │   184 │    16 │ 0.08      │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          5 │  1615 │   139 │ 0.0792474 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          6 │  1180 │    91 │ 0.0715972 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          9 │   304 │    21 │ 0.0646154 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          7 │   677 │    46 │ 0.0636238 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│          8 │   504 │    31 │ 0.0579439 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         10 │   250 │    14 │ 0.0530303 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         15 │    79 │     4 │ 0.0481928 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         13 │   126 │     6 │ 0.0454545 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         14 │    89 │     4 │ 0.0430108 │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         12 │   150 │     4 │ 0.025974  │\n",
      "├────────────┼───────┼───────┼───────────┤\n",
      "│         16 │    76 │     2 │ 0.025641  │\n",
      "╘════════════╧═══════╧═══════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "dfc = pd.crosstab(data1.y, [data1.campaign]).T\n",
    "dfc['success']=dfc.yes/(dfc.no+dfc.yes) \n",
    "dfc = dfc.sort_values(by=['success'], ascending=False)\n",
    "print('Rank considering only campaign\\'s calls')\n",
    "print(tabulate(dfc, tablefmt='fancy_grid', headers='keys', numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Detail of winner of total calls (tt_contacts=1) \n",
      "╒══════════════╤═══════╤═══════╤═══════════╕\n",
      "│ 9/previous   │    no │   yes │   success │\n",
      "╞══════════════╪═══════╪═══════╪═══════════╡\n",
      "│ (1, 0)       │ 12006 │  1523 │  0.112573 │\n",
      "╘══════════════╧═══════╧═══════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "champions_campaign = data1[data1.tt_contacts==1]\n",
    "ctchampions_campaign = pd.crosstab(champions_campaign.y, [champions_campaign.tt_contacts,champions_campaign.previous]).T\n",
    "ctchampions_campaign['perc'] = ctchampions_campaign.yes/(ctchampions_campaign.yes+ctchampions_campaign.no)\n",
    "ctchampions_campaign = ctchampions_campaign.sort_values(by=['perc'], ascending=False)\n",
    "ctchampions_campaign.index.names=['a','b']\n",
    "\n",
    "print('Detail of winner of total calls (tt_contacts=1) ')\n",
    "print(tabulate(ctchampions_campaign, tablefmt='fancy_grid', headers=['9/previous', 'no', 'yes', 'success'], numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rank considering total calls\n",
      "╒═══════════════╤═══════╤═══════╤═══════════╕\n",
      "│   tt_contacts │    no │   yes │   success │\n",
      "╞═══════════════╪═══════╪═══════╪═══════════╡\n",
      "│             9 │   460 │    86 │ 0.157509  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            11 │   274 │    47 │ 0.146417  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             5 │  2171 │   343 │ 0.136436  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             6 │  1576 │   243 │ 0.13359   │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             7 │   953 │   146 │ 0.132848  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            10 │   388 │    58 │ 0.130045  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             4 │  3850 │   569 │ 0.128762  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             3 │  5644 │   805 │ 0.124826  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             8 │   730 │   104 │ 0.1247    │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            12 │   219 │    28 │ 0.11336   │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             1 │ 12006 │  1523 │ 0.112573  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            15 │   119 │    15 │ 0.11194   │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            14 │   129 │    16 │ 0.110345  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│             2 │ 10513 │  1247 │ 0.106037  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            13 │   187 │    22 │ 0.105263  │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            16 │   107 │     8 │ 0.0695652 │\n",
      "├───────────────┼───────┼───────┼───────────┤\n",
      "│            17 │    91 │     6 │ 0.0618557 │\n",
      "╘═══════════════╧═══════╧═══════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "dft = pd.crosstab(data1.y, [data1.tt_contacts]).T\n",
    "dft['success']=dft.yes/(dft.no+dft.yes) \n",
    "dft = dft.sort_values(by=['success'], ascending=False)\n",
    "print('Rank considering total calls')\n",
    "print(tabulate(dft, tablefmt='fancy_grid', headers='keys', numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Detail of winner of total calls (tt_contacts=9) \n",
      "╒══════════════╤══════╤═══════╤═══════════╕\n",
      "│ 9/previous   │   no │   yes │   success │\n",
      "╞══════════════╪══════╪═══════╪═══════════╡\n",
      "│ (9, 6)       │   24 │    16 │ 0.4       │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 8)       │   34 │    19 │ 0.358491  │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 7)       │   37 │    15 │ 0.288462  │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 4)       │   17 │     6 │ 0.26087   │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 5)       │   24 │     5 │ 0.172414  │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 2)       │   12 │     2 │ 0.142857  │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 3)       │   18 │     3 │ 0.142857  │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 1)       │   10 │     1 │ 0.0909091 │\n",
      "├──────────────┼──────┼───────┼───────────┤\n",
      "│ (9, 0)       │  284 │    19 │ 0.0627063 │\n",
      "╘══════════════╧══════╧═══════╧═══════════╛\n"
     ]
    }
   ],
   "source": [
    "champions = data1[data1.tt_contacts==9]\n",
    "ctchampions = pd.crosstab(champions.y, [champions.tt_contacts,champions.previous]).T\n",
    "ctchampions['perc'] = ctchampions.yes/(ctchampions.yes+ctchampions.no)\n",
    "ctchampions = ctchampions.sort_values(by=['perc'], ascending=False)\n",
    "ctchampions.index.names=['a','b']\n",
    "\n",
    "print('Detail of winner of total calls (tt_contacts=9) ')\n",
    "print(tabulate(ctchampions, tablefmt='fancy_grid', headers=['9/previous', 'no', 'yes', 'success'], numalign=\"decimal\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "- The Pearson's chi-squared statistic test was avoided by  there were frequencies under 5, then I was used 2 tests:  Log-likelihood ratio (G-test) and Freeman-Tukey statistic, in both results, is possible to notice that we can reject the null hypothesis, i.e. there is a correlation between the number of contacts and the success of campaign.\n",
    "\n",
    "- The best number of total calls was 9 with 15,7% of success, inside this result (9 class) the best result is those with 6 calls on previous\n",
    "\n",
    "- Curiously, all records with only one call in the campaign had not any call on previous."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}