data-preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Document dependencies\n",
    "Skip this section. To document the dependencies, execute the following cells after executing all relevant import statements."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install --user watermark"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "CPython 3.7.4\n",
      "IPython 7.8.0\n",
      "\n",
      "numpy  1.18.5\n",
      "pandas 0.25.1\n",
      "\n"
     ]
    }
   ],
   "source": [
    "%load_ext watermark\n",
    "print()\n",
    "%watermark -v\n",
    "print()\n",
    "%watermark -iv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pre-processing\n",
    "\n",
    "This notebook contains the pre-processing of Annie's EbolaDatabase for the purposes of the semester project. These purposes are:\n",
    "- To replicate some of the prediction models described in her 2017 papers, in the centralized as well as various simulated distributed ML settings.\n",
    "- To define other models, if they can be useful to benchmark model personalisation / data selection schemes.\n",
    "\n",
    "# 1. Read the raw EbolaDatabase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"../data/private/0_raw_EbolaDatabase.csv\")\n",
    "df_in = df.copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let me show you something: **study_id 125 is missing**. That's not really problematic, but _do not_ use `study_id` as index!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "123    124\n",
       "124    126\n",
       "125    127\n",
       "126    128\n",
       "127    129\n",
       "Name: study_id, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_in.study_id[123:128]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>study_id</th>\n",
       "      <th>evd</th>\n",
       "      <th>evd_ct</th>\n",
       "      <th>malaria</th>\n",
       "      <th>death</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>quarantine</th>\n",
       "      <th>days_admitted</th>\n",
       "      <th>referral_time</th>\n",
       "      <th>date_admis</th>\n",
       "      <th>temp_triage</th>\n",
       "      <th>pyrexia_triage</th>\n",
       "      <th>anuria_triage</th>\n",
       "      <th>dyspnea_triage</th>\n",
       "      <th>disorientation_triage</th>\n",
       "      <th>headache_triage</th>\n",
       "      <th>hiccups_triage</th>\n",
       "      <th>myalgia_triage</th>\n",
       "      <th>rash_triage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [study_id, evd, evd_ct, malaria, death, sex, age, quarantine, days_admitted, referral_time, date_admis, temp_triage, pyrexia_triage, anuria_triage, dyspnea_triage, disorientation_triage, headache_triage, hiccups_triage, myalgia_triage, rash_triage]\n",
       "Index: []"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.iloc[:0, :20] # DO NOT PUSH A COMMIT WHERE ANY LINE IS VISIBLE!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Missing values\n",
    "\n",
    "Missing values must be dealt with, in one way or another, for all features that will be used in machine learning models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(575, 67)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>study_id</th>\n",
       "      <th>evd</th>\n",
       "      <th>evd_ct</th>\n",
       "      <th>malaria</th>\n",
       "      <th>death</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>quarantine</th>\n",
       "      <th>days_admitted</th>\n",
       "      <th>referral_time</th>\n",
       "      <th>...</th>\n",
       "      <th>hemoptysis_after</th>\n",
       "      <th>hematochezia_after</th>\n",
       "      <th>hematuria_after</th>\n",
       "      <th>gyne_hemorrhage_after</th>\n",
       "      <th>num_hemorrhage_sites_after</th>\n",
       "      <th>dehydration_after</th>\n",
       "      <th>hepatomegaly_after</th>\n",
       "      <th>jaundice_after</th>\n",
       "      <th>edema_after</th>\n",
       "      <th>splenomegaly_after</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 67 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   study_id    evd  evd_ct  malaria  death    sex    age  quarantine  \\\n",
       "0     False  False    True     True  False  False  False        True   \n",
       "\n",
       "   days_admitted  referral_time  ...  hemoptysis_after  hematochezia_after  \\\n",
       "0          False           True  ...             False               False   \n",
       "\n",
       "   hematuria_after  gyne_hemorrhage_after  num_hemorrhage_sites_after  \\\n",
       "0            False                  False                       False   \n",
       "\n",
       "   dehydration_after  hepatomegaly_after  jaundice_after  edema_after  \\\n",
       "0              False               False           False        False   \n",
       "\n",
       "   splenomegaly_after  \n",
       "0               False  \n",
       "\n",
       "[1 rows x 67 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df.shape)\n",
    "# Show which columns contain any NA values:\n",
    "df.isna().any().to_frame().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['study_id', 'evd', 'evd_ct', 'malaria', 'death', 'sex', 'age',\n",
       "       'quarantine', 'days_admitted', 'referral_time', 'date_admis',\n",
       "       'temp_triage', 'pyrexia_triage', 'anuria_triage', 'dyspnea_triage',\n",
       "       'disorientation_triage', 'headache_triage', 'hiccups_triage',\n",
       "       'myalgia_triage', 'rash_triage', 'dysphagia_triage', 'asthenia_triage',\n",
       "       'abdomnial_triage', 'anorexia_triage', 'diarrhea_triage',\n",
       "       'diarrhea_score_triage', 'vomiting_triage', 'hemorrhage_triage',\n",
       "       'orl_hemorrhage_triage', 'conjonctivitis_triage', 'hemoptysis_triage',\n",
       "       'hematochezia_triage', 'hematuria_triage', 'gynec_hemorrhage_triage',\n",
       "       'num_hemorrhages_triage', 'mult_hemorrhage_triage',\n",
       "       'dehydration_triage', 'hepatomegaly_triage', 'jaundice_triage',\n",
       "       'edema_triage', 'splenomegaly_triage', 'anuria_after', 'dyspnea_after',\n",
       "       'disorientation_after', 'headache_after', 'hiccups_after',\n",
       "       'myalgia_after', 'rash_after', 'dysphagia_after', 'asthenia_after',\n",
       "       'abdominal_after', 'anorexia_after', 'diarrhea_score_after',\n",
       "       'vomiting_after', 'hemorrhage_after', 'orl_hemorrhage_after',\n",
       "       'conjunctivitis_after', 'hemoptysis_after', 'hematochezia_after',\n",
       "       'hematuria_after', 'gyne_hemorrhage_after',\n",
       "       'num_hemorrhage_sites_after', 'dehydration_after', 'hepatomegaly_after',\n",
       "       'jaundice_after', 'edema_after', 'splenomegaly_after'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "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>evd_ct</th>\n",
       "      <th>malaria</th>\n",
       "      <th>quarantine</th>\n",
       "      <th>referral_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>431</td>\n",
       "      <td>32</td>\n",
       "      <td>24</td>\n",
       "      <td>87</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   evd_ct  malaria  quarantine  referral_time\n",
       "0     431       32          24             87"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.columns[df.isna().any()]].isna().sum().to_frame().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "evd_ct is available for 144 of the 158 EVD+ patients.\n"
     ]
    }
   ],
   "source": [
    "print(\"evd_ct is available for {ct} of the {pos} EVD+ patients.\".format(\n",
    "        ct=len(df) - df['evd_ct'].isna().sum(),\n",
    "        pos=df['evd'].sum()\n",
    "))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first four columns are the most obvious targets (`evd`, `evd_ct`, `malaria` and `death`).\n",
    "The first three could be particularly useful in predicting `death`.\n",
    "\n",
    "Every patient admitted to the ETC was tested for Ebola (column `evd`).\n",
    "`evd_ct` (inversely proportional to the viral load) is missing for all EVD- patients and for 14 EVD+ patients.\n",
    "Malaria test results are missing for 32 patients."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.1. Referral time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Annie's paper mentions 20 missing values for the referral time, which were tested to be likely missing completely at random (MCAR).\n",
    "She transforms this into the binary variable $\\mathbb{1}\\left\\{ x \\in [4, 9] \\right\\}$ ($x \\triangleq$ referral time in days).\n",
    "Let's plot the proportion of EVD(+) as a function of referral time, along with confidence intervals (a Wilson score interval obtained by assuming that the EVD diagnosis follows a Bernoulli distribution whose probability depends on the referral time)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>evd</th>\n",
       "      <th>referral_time</th>\n",
       "      <th>total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>n</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>06</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>01</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>n</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   evd referral_time  total\n",
       "0    1            12      1\n",
       "1    1             n      1\n",
       "2    1            06      1\n",
       "3    1            01      1\n",
       "4    1             n      1"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_ref = df['evd'].to_frame()\n",
    "df_ref['referral_time'] = df['referral_time'].astype(str).apply(\n",
    "        lambda s: s[:-2] if ('na' in s.lower() or len(s)==4) else '0'+s[:-2]\n",
    ")\n",
    "df_ref['total'] = 1\n",
    "df_ref.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'number of patients')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_ref = df_ref.groupby('referral_time').sum()\n",
    "df_ref.plot(y='total', xticks=range(len(df_ref)), figsize=(12,6), grid=True, ylim=(0,100), yticks=[0,1]+list(range(10,101, 10)), title='Histogram of referral times')\n",
    "plt.ylabel(\"number of patients\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On second thought, these intervals (and also the proportion of EVD(+) patients in each `referral_time` category) will be near meaningless, considering the small number of samples per category.\n",
    "Instead, we follow Annie in replacing the referral_time by a binary variable (cf. previous Markdown cell).\n",
    "In this new binary feature, we replace `NaN`s by the proportion of `1`s among available values.\n",
    "With this, the feature is no longer truly binary, but can instead be interpreted as an a-priori probability of having a referral time in between 4 and 9 days.\n",
    "This probability is exactly 1 or 0 for all patients whose referral time is known."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  1,   4,   9,  16,  52,  75,  81,  83,  93,  98, 102, 125, 132,\n",
       "       133, 138, 141, 143, 144, 150, 151, 152, 154, 156, 167, 168, 172,\n",
       "       183, 192, 211, 217, 218, 233, 236, 237, 238, 244, 245, 247, 248,\n",
       "       253, 278, 279, 290, 300, 301, 307, 311, 317, 322, 327, 335, 336,\n",
       "       355, 357, 360, 363, 368, 395, 396, 398, 399, 402, 410, 411, 412,\n",
       "       416, 421, 435, 441, 449, 451, 453, 454, 460, 464, 478, 483, 485,\n",
       "       487, 492, 502, 515, 518, 524, 542, 545, 561])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['referral(4-9)'] = ((df['referral_time'] <= 9) & (df['referral_time'] >= 4)).astype(np.float32)\n",
    "np.where(np.isnan(df['referral_time']))[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>122</th>\n",
       "      <th>123</th>\n",
       "      <th>124</th>\n",
       "      <th>125</th>\n",
       "      <th>126</th>\n",
       "      <th>127</th>\n",
       "      <th>128</th>\n",
       "      <th>129</th>\n",
       "      <th>130</th>\n",
       "      <th>131</th>\n",
       "      <th>132</th>\n",
       "      <th>133</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>study_id</td>\n",
       "      <td>123.0</td>\n",
       "      <td>124.0</td>\n",
       "      <td>126.0</td>\n",
       "      <td>127.0</td>\n",
       "      <td>128.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>132.0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>134.0</td>\n",
       "      <td>135.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>referral_time</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>referral(4-9)</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 122    123    124    125    126    127    128    129    130  \\\n",
       "study_id       123.0  124.0  126.0  127.0  128.0  129.0  130.0  131.0  132.0   \n",
       "referral_time    5.0    5.0    6.0    NaN    3.0    1.0    6.0    3.0    5.0   \n",
       "referral(4-9)    1.0    1.0    1.0    NaN    0.0    0.0    1.0    0.0    1.0   \n",
       "\n",
       "                 131    132    133  \n",
       "study_id       133.0  134.0  135.0  \n",
       "referral_time    4.0    NaN    NaN  \n",
       "referral(4-9)    1.0    NaN    NaN  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[ np.where(np.isnan(df['referral_time']))[0] , 'referral(4-9)'] = float('NaN')\n",
    "df.loc[122:133, [ 'study_id', 'referral_time', 'referral(4-9)']].T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "False\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>122</th>\n",
       "      <th>123</th>\n",
       "      <th>124</th>\n",
       "      <th>125</th>\n",
       "      <th>126</th>\n",
       "      <th>127</th>\n",
       "      <th>128</th>\n",
       "      <th>129</th>\n",
       "      <th>130</th>\n",
       "      <th>131</th>\n",
       "      <th>132</th>\n",
       "      <th>133</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>study_id</td>\n",
       "      <td>123.0</td>\n",
       "      <td>124.0</td>\n",
       "      <td>126.0</td>\n",
       "      <td>127.000000</td>\n",
       "      <td>128.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>132.0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>134.000000</td>\n",
       "      <td>135.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>referral_time</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>referral(4-9)</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.364754</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.364754</td>\n",
       "      <td>0.364754</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 122    123    124         125    126    127    128    129  \\\n",
       "study_id       123.0  124.0  126.0  127.000000  128.0  129.0  130.0  131.0   \n",
       "referral_time    5.0    5.0    6.0         NaN    3.0    1.0    6.0    3.0   \n",
       "referral(4-9)    1.0    1.0    1.0    0.364754    0.0    0.0    1.0    0.0   \n",
       "\n",
       "                 130    131         132         133  \n",
       "study_id       132.0  133.0  134.000000  135.000000  \n",
       "referral_time    5.0    4.0         NaN         NaN  \n",
       "referral(4-9)    1.0    1.0    0.364754    0.364754  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[ np.where(np.isnan(df['referral_time']))[0] , 'referral(4-9)'] = df['referral(4-9)'].mean()\n",
    "print(df['referral(4-9)'].isna().any())\n",
    "df.loc[122:133, [ 'study_id', 'referral_time', 'referral(4-9)']].T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2. Quarantine\n",
    "\n",
    "The single most useful feature in the predictive models described in Annie's papers is whether or not the patient had had contact with a confirmed EVD patient.\n",
    "All those known to have had contact with confirmed EVD patients, underwent a quarantine (often together in groups).\n",
    "Thus, the `quarantine` column can be interpreted as \"We know that this patient has had contact with a confirmed EVD patient\".\n",
    "\n",
    "Unfortunately, 24 patients do not have an entry in this column.\n",
    "In principle, this means _we do not know whether this patient has had contact with a confirmed EVD patient_.\n",
    "What we do with these missing values largely depends on how the data for this feature was gathered.\n",
    "\n",
    "Below, we perform a visual inspection of how the missing values are distributed:\n",
    "<!-- Even if the values are missing completely at random, the best way to fill missing values depends on which scenario we would rather believe:\n",
    "- Of course, if the values are not missing completely at random, it might be useful to include a binary feature for whether or not the `quarantine` value is missing.\n",
    "- Else, if we believe that nearly all individuals who have had contact with a confirmed EVD patient really underwent quarantine, then missing values should be replaced by the column mean.\n",
    "- If, on the contrary, we believe that many (say, more than half) of the patients who have had contact with a confirmed EVD patient, are never found -->"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>study_id</th>\n",
       "      <th>evd</th>\n",
       "      <th>evd_ct</th>\n",
       "      <th>malaria</th>\n",
       "      <th>death</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>quarantine</th>\n",
       "      <th>days_admitted</th>\n",
       "      <th>referral_time</th>\n",
       "      <th>...</th>\n",
       "      <th>hematochezia_after</th>\n",
       "      <th>hematuria_after</th>\n",
       "      <th>gyne_hemorrhage_after</th>\n",
       "      <th>num_hemorrhage_sites_after</th>\n",
       "      <th>dehydration_after</th>\n",
       "      <th>hepatomegaly_after</th>\n",
       "      <th>jaundice_after</th>\n",
       "      <th>edema_after</th>\n",
       "      <th>splenomegaly_after</th>\n",
       "      <th>referral(4-9)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 68 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [study_id, evd, evd_ct, malaria, death, sex, age, quarantine, days_admitted, referral_time, date_admis, temp_triage, pyrexia_triage, anuria_triage, dyspnea_triage, disorientation_triage, headache_triage, hiccups_triage, myalgia_triage, rash_triage, dysphagia_triage, asthenia_triage, abdomnial_triage, anorexia_triage, diarrhea_triage, diarrhea_score_triage, vomiting_triage, hemorrhage_triage, orl_hemorrhage_triage, conjonctivitis_triage, hemoptysis_triage, hematochezia_triage, hematuria_triage, gynec_hemorrhage_triage, num_hemorrhages_triage, mult_hemorrhage_triage, dehydration_triage, hepatomegaly_triage, jaundice_triage, edema_triage, splenomegaly_triage, anuria_after, dyspnea_after, disorientation_after, headache_after, hiccups_after, myalgia_after, rash_after, dysphagia_after, asthenia_after, abdominal_after, anorexia_after, diarrhea_score_after, vomiting_after, hemorrhage_after, orl_hemorrhage_after, conjunctivitis_after, hemoptysis_after, hematochezia_after, hematuria_after, gyne_hemorrhage_after, num_hemorrhage_sites_after, dehydration_after, hepatomegaly_after, jaundice_after, edema_after, splenomegaly_after, referral(4-9)]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 68 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_quarantine_missing = df.iloc[ np.where(np.isnan(df['quarantine']))[0] , :].copy()\n",
    "print(len(df_quarantine_missing))\n",
    "df_quarantine_missing.head(0) # DO NOT PUSH A COMMIT WHERE ANY LINE IS VISIBLE!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "415\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_quarantine_0 = df.loc[ df['quarantine'] == 0 , :].copy()\n",
    "print(len(df_quarantine_0))\n",
    "df_quarantine_0.quarantine.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "136\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_quarantine_1 = df.loc[ df['quarantine'] == 1 , :].copy()\n",
    "print(len(df_quarantine_1))\n",
    "df_quarantine_1.quarantine.min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "non_binary_features = {'study_id', 'evd_ct', 'age', 'days_admitted', 'referral_time', 'date_admis', 'temp_triage', 'num_hemorrhages_triage', 'num_hemorrhage_sites_after'}\n",
    "\n",
    "def compare_means(dfs, labels):\n",
    "    dfs_binary = [df.drop(labels=non_binary_features.intersection(df.columns), axis=1) for df in dfs]\n",
    "    \n",
    "    interesting_non_binary = {'age', 'date_admis', 'study_id'}\n",
    "    non_interesting_features = non_binary_features - interesting_non_binary\n",
    "    if not all((set(df.columns) - non_interesting_features) == (set(dfs[0].columns) - non_interesting_features) for df in dfs):\n",
    "        raise ValueError(\"The passed DataFrames do not have the same column labels.\")\n",
    "    if not all((df.columns == dfs_binary[0].columns).all() for df in dfs_binary):\n",
    "        raise ValueError(\"The column labels of the passed DataFrames do not follow the same ordering.\")\n",
    "        \n",
    "    fig, ax = plt.subplots(figsize=(25,7))\n",
    "    xlabels = dfs_binary[0].columns.to_list()\n",
    "    x = np.arange(len(xlabels))\n",
    "    width = 0.75/len(dfs)\n",
    "    for i, df in enumerate(dfs_binary):\n",
    "        ax.bar(x + (i - len(dfs)/2.)*width, df.mean(axis=0), width, label=labels[i])\n",
    "        \n",
    "    ax.legend()\n",
    "    ax.set_xticks(x)\n",
    "    ax.set_xticklabels(xlabels, rotation=90)\n",
    "    ax.set_xlim((-1, len(xlabels)))\n",
    "    ax.set_ylim((0,1))\n",
    "    ax.set_ylabel(\"Proportion of `True`\")\n",
    "    plt.show()\n",
    "    \n",
    "    # The non-binary features `age` and `date_admis` could also revel, non-random missingness:\n",
    "    fig, axs = plt.subplots(ncols=len(interesting_non_binary), figsize=(25,7))\n",
    "    for ax, col_name in zip(axs.flat, interesting_non_binary):\n",
    "        for df, label in zip(dfs, labels):\n",
    "            ax.hist(df[col_name], label=label, histtype='step', density=True)\n",
    "        ax.legend()\n",
    "        ax.set_title(col_name + ' distribution')\n",
    "        ax.set_xlabel(\"Age [years]\" if col_name=='age' else \"Month of 2015\")\n",
    "        ax.set_ylabel(\"Density\")\n",
    "    plt.show()\n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x504 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ax = df_quarantine.drop(labels='study_id', axis=1).mean(axis=0).plot(kind='bar', figsize=(20, 6), grid=True, ylim=(0,1))\n",
    "compare_means([df, df_quarantine_missing, df_quarantine_0, df_quarantine_1], ['entire population', '`quarantine` missing', '`quarantine` == 0', 'quarantine == 1'])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the bar plot, we identify some features that correlate relatively strongly with quarantine (large discrepancy between `quarantine == 0` and `quarantine == 1`):\n",
    "- evd\n",
    "- death\n",
    "- \\* anuria\n",
    "- \\* dyspnea (only at triage)\n",
    "- diarrhea\n",
    "- vomiting\n",
    "- oral hemorrhage\n",
    "- \\* conjunctivitis\n",
    "\n",
    "We confirm our previous intuition that `NaN` values should not be replaced with `1`. While some features (highlighted with \\*) seem to suggest that the `NaN` population is actually more alike to the `0` population, than to the entire population, this claim can not be made with any reasonable degree of confidence based on the low number of 24 samples with missing `quarantine` values.\n",
    "In a situation like this, we would generally replace the missing values by the column mean due to the low number of missing samples.\n",
    "As `quarantine` was shown to be the most predictive value, however, we might be justified to create a separate binary variable indicating whether the `quarantine` value was missing.\n",
    "This has the effect that the model learns which value would have been most appropriate to replace missing values. This approach also captures the effect of any confounding variable that could be responsible for the missing values (and potentially affect the outcome for the patients), thus making it well-suited to deal with non-MCAR (missing _not_ completely at random) data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24\n"
     ]
    }
   ],
   "source": [
    "assert df['quarantine'].isna().sum() > 0\n",
    "df['quarantine_missing'] = df['quarantine'].isna().astype('int')\n",
    "print(df['quarantine_missing'].sum())\n",
    "df.loc[ np.where(np.isnan(df['quarantine']))[0] , 'quarantine'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>study_id</th>\n",
       "      <th>evd</th>\n",
       "      <th>evd_ct</th>\n",
       "      <th>malaria</th>\n",
       "      <th>death</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>quarantine</th>\n",
       "      <th>days_admitted</th>\n",
       "      <th>referral_time</th>\n",
       "      <th>...</th>\n",
       "      <th>hematuria_after</th>\n",
       "      <th>gyne_hemorrhage_after</th>\n",
       "      <th>num_hemorrhage_sites_after</th>\n",
       "      <th>dehydration_after</th>\n",
       "      <th>hepatomegaly_after</th>\n",
       "      <th>jaundice_after</th>\n",
       "      <th>edema_after</th>\n",
       "      <th>splenomegaly_after</th>\n",
       "      <th>referral(4-9)</th>\n",
       "      <th>quarantine_missing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 69 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [study_id, evd, evd_ct, malaria, death, sex, age, quarantine, days_admitted, referral_time, date_admis, temp_triage, pyrexia_triage, anuria_triage, dyspnea_triage, disorientation_triage, headache_triage, hiccups_triage, myalgia_triage, rash_triage, dysphagia_triage, asthenia_triage, abdomnial_triage, anorexia_triage, diarrhea_triage, diarrhea_score_triage, vomiting_triage, hemorrhage_triage, orl_hemorrhage_triage, conjonctivitis_triage, hemoptysis_triage, hematochezia_triage, hematuria_triage, gynec_hemorrhage_triage, num_hemorrhages_triage, mult_hemorrhage_triage, dehydration_triage, hepatomegaly_triage, jaundice_triage, edema_triage, splenomegaly_triage, anuria_after, dyspnea_after, disorientation_after, headache_after, hiccups_after, myalgia_after, rash_after, dysphagia_after, asthenia_after, abdominal_after, anorexia_after, diarrhea_score_after, vomiting_after, hemorrhage_after, orl_hemorrhage_after, conjunctivitis_after, hemoptysis_after, hematochezia_after, hematuria_after, gyne_hemorrhage_after, num_hemorrhage_sites_after, dehydration_after, hepatomegaly_after, jaundice_after, edema_after, splenomegaly_after, referral(4-9), quarantine_missing]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 69 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(0) # DO NOT PUSH A COMMIT WHERE ANY LINE IS VISIBLE!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Saving the result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('../data/private/1_TreatMissing_EbolaDatabase.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f6b966de610>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.plot(y='age', kind='hist', bins=21)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Many columns are binary. Let's count the positives (meaningless for non-binary features):\n",
    "df.sum().to_frame().plot(kind='bar', figsize=(15,7))\n",
    "plt.gca().set_ylim((0, len(df)))\n",
    "hline_limit = 75\n",
    "plt.gca().axhline(y=hline_limit, color='r')\n",
    "plt.gca().axhline(y=len(df) - hline_limit, color='r')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One possibility for dealing with continuous or ordered features for which values are missing not completely at random (not MCAR), at least for ridge or logistic regression:\n",
    "\n",
    "1. Standardize the features if you're going to use regularization (and why wouldn't you)\n",
    "2. Add a one-hot encoded feature \"feature_XY_is_na\" (this feature should be non-zero when the other feature _is_ missing).\n",
    "3. Replace the na values by 0.\n",
    "\n",
    "Result: one weight learns what to contribute to the model when the feature is na, and another learns what to do when it is available. This could, for instance, result in different intercepts for people with vs without missing value, which makes sense if the values are missing for some systematic reason. It's kind of a backwards logic of \"if the Doctor didn't deem it useful to perform this test, that's probably because the person was fine\" or something.\n",
    "\n",
    "Main downside: If the number of missing values is low, "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}