hcpn-network-NS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Network Analysis of the Primary Healthcare Facility Referral System in the Philippines\n",
    "##### By: Elijah Justin Medina and Jomilynn Rebanal\n",
    "\n",
    "This project aims to analyze to referral system among health units in the Philippines in order to identify the characteristics and consequently, the robustness of this system. The main data used in this analysis are geospatial coordinates of different health units in the network &mdash; barangay health stations (BHS), rural health units (RHU), and hospitals. The network is supplemented with bed capacity information of the different health units. The aforementioned data was obtained from the DOH Data Collect app v2.1 and the National Health Facility Registry."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Preliminaries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-03T13:52:37.860311Z",
     "start_time": "2020-08-03T13:52:37.855324Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np \n",
    "from glob import glob \n",
    "import sys\n",
    "import locale\n",
    "from geopy.distance import vincenty\n",
    "import warnings\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "np.set_printoptions(threshold=sys.maxsize, precision=2) \n",
    "pd.set_option('float_format', '{:,.2f}'.format)\n",
    "pd.set_option('display.max_rows', 1000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Processing and Exploratory Data Analysis\n",
    "Before performing the analysis, the data must be cleaned to standardize text and minimize effect of errors (e.g. typographical, encoding, outliers). The goal of this processing is to prepare the data for network analysis."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Merging of multiple, separated files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Geographic Coordinates (Health Facility names and coordinates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:05.277261Z",
     "start_time": "2020-03-10T12:49:05.269308Z"
    }
   },
   "outputs": [],
   "source": [
    "files = glob('Geographic Coordinates/*.xlsx')\n",
    "cols = ['Health Facility Code Short', 'Facility Name', 'Health Facility Type',\n",
    "        'Region Name                                             ', \n",
    "        'Province Name ', 'City/Municipality Name', 'Barangay Name', \n",
    "        'Latitude', 'Longitude', 'Service Capability', 'Licensing Status', \n",
    "        'Bed Capacity']\n",
    "\n",
    "HF_list = pd.DataFrame()\n",
    "        \n",
    "for f in files:\n",
    "    data = pd.read_excel(f, usecols=cols)\n",
    "    HF_list = HF_list.append(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:05.776642Z",
     "start_time": "2020-03-10T12:49:05.768638Z"
    }
   },
   "outputs": [],
   "source": [
    "HF_list.isnull().sum() # Verify mismatched fields across different excel files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:06.268012Z",
     "start_time": "2020-03-10T12:49:06.259985Z"
    }
   },
   "outputs": [],
   "source": [
    "HF_list.columns = ['SHORT_HFCODE', 'HF_NAME', 'HF_TYPE', 'REGION', 'PROVINCE', \n",
    "                  'MUNI_CITY', 'BRGY', 'LAT', 'LONG', 'SERVICE_CAP', \n",
    "                  'LICENSING', 'BED_CAP']\n",
    "\n",
    "str_cols = ['HF_NAME', 'HF_TYPE', 'REGION', 'PROVINCE', 'MUNI_CITY', 'BRGY', \n",
    "            'SERVICE_CAP', 'LICENSING', 'BED_CAP']\n",
    "\n",
    "HF_list[str_cols] = HF_list[str_cols].fillna('UNKNOWN').apply(lambda x: x.str.upper().str.strip())\n",
    "HF_list['SHORT_HFCODE'] = HF_list['SHORT_HFCODE'].astype(int)\n",
    "\n",
    "HF_list.to_excel('cleaned/HFList_cleaned.xlsx') #Store the combined dataframe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Rural Health Unit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:07.746989Z",
     "start_time": "2020-03-10T12:49:07.742995Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu = pd.read_excel('rhu2018.xlsx', sheet_name='MAIN', na_values='None')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:08.242363Z",
     "start_time": "2020-03-10T12:49:08.234372Z"
    }
   },
   "outputs": [],
   "source": [
    "str_cols = ['HF_NAME', 'REGION', 'PROVINCE', 'MUNI_CITY', 'BRGY',\n",
    "            'STREET_NAME', 'BUILDING', 'FACILITY_HEAD', 'DETACHED', 'BRGYS', \n",
    "            'SDN', 'SDN_NAME', 'REF1_NAME', 'REF1_SAMEPROV', \n",
    "            'REF1_REF1A', 'REF1A_SAMEPROV', 'REF2_NAME', 'REF3_NAME',  \n",
    "            'AMB_ACCESS', 'AMB_OWN', 'PHIC_ACC', 'PHIC_PACKAGES', 'PHIC_MCP', \n",
    "            'PHIC_PCB1', 'PHIC_MALARIA', 'PHIC_TBDOTS', 'PHIC_ABTC', \n",
    "            'PHIC_NCP', 'PHIC_OTH']\n",
    "code_cols = ['id', 'REF1_CODE', 'REF2_CODE', 'REF3_CODE']\n",
    "float_cols = ['REF1_DIST', 'REF1_TRAVEL', 'REF2_DIST', 'REF2_TRAVEL',\n",
    "             'REF3_DIST', 'REF3_TRAVEL']\n",
    "# int_cols = ['id', 'BHS_COUNT','CATCHMENT', 'REF1_CODE', 'REF2_CODE', \n",
    "#             'REF3_CODE',  'MD_NO', 'MD_AUG', 'MD_TOTAL','MD_FT', 'MD_PT', \n",
    "#             'MD_VISIT', 'RN_NO', 'RN_AUG', 'RN_TOTAL', 'RN_FT', 'RN_PT', \n",
    "#             'RN_VISIT', 'MW_NO', 'MW_AUG', 'MW_TOTAL', 'MW_FT', 'MW_PT', \n",
    "#             'MW_VISIT']\n",
    "\n",
    "rhu[str_cols] = rhu[str_cols].apply(lambda x: x.str.upper().str.strip())\n",
    "rhu[code_cols] = rhu[code_cols].fillna(0).astype(int)\n",
    "rhu[float_cols] = rhu[float_cols].astype(float)\n",
    "\n",
    "rhu[str_cols] = rhu[str_cols].fillna('UNKNOWN')\n",
    "\n",
    "# Extract short code for merging with the Geographic Coordinates files\n",
    "rhu['SHORT_HFCODE'] = rhu['HF_CODE'].apply(lambda x: int(x[-6:]))\n",
    "\n",
    "rhu.to_excel('cleaned/rhu_cleaned.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-10T12:49:09.235805Z",
     "start_time": "2020-03-10T12:49:09.227809Z"
    }
   },
   "source": [
    "### Impute missing information from other tables\n",
    "As the data is being processed from different tables, the other tables can be used to fill some missing information. Aside from imputing missing information, coordinates outside the Philippines are identified."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:40.043551Z",
     "start_time": "2020-08-10T18:34:40.039563Z"
    }
   },
   "outputs": [],
   "source": [
    "# Bounding box for the Philippines (manually extracted from Google).\n",
    "long_min, lat_min, long_max, lat_max = (117.17427453, 5.58100332277, 126.537423944, 18.5052273625)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:47.943427Z",
     "start_time": "2020-08-10T18:34:40.593814Z"
    }
   },
   "outputs": [],
   "source": [
    "HF_list = pd.read_excel('cleaned/HFList_cleaned.xlsx')\n",
    "\n",
    "# Groupby the data to account for duplicate names for different codes\n",
    "HF_dict = HF_list[['HF_NAME', 'SHORT_HFCODE']].groupby('HF_NAME')['SHORT_HFCODE'].apply(set).to_dict()\n",
    "latlong_dict = HF_list[['SHORT_HFCODE', 'LAT', 'LONG']].set_index('SHORT_HFCODE').to_dict()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### RHU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:53.260437Z",
     "start_time": "2020-08-10T18:34:52.040575Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu = pd.read_excel('cleaned/rhu_cleaned.xlsx')\n",
    "\n",
    "# Create copies of the dataframe for later use\n",
    "rhu2 = rhu.copy()\n",
    "rhu3 = rhu.copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Fill missing REF1 Codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:54.770202Z",
     "start_time": "2020-08-10T18:34:54.708368Z"
    }
   },
   "outputs": [],
   "source": [
    "cols = ['id', 'HF_CODE', 'SHORT_HFCODE', 'HF_NAME', 'REGION', 'PROVINCE', 'MUNI_CITY', 'BRGY',\n",
    "            'STREET_NAME', 'BUILDING', 'FACILITY_HEAD', 'DETACHED', 'BRGYS', \n",
    "            'SDN', 'SDN_NAME', 'REF1_NAME', 'REF1_SAMEPROV', \n",
    "            'REF1_REF1A', 'REF1A_SAMEPROV', 'REF2_NAME', 'REF3_NAME',  \n",
    "            'AMB_ACCESS', 'AMB_OWN', 'PHIC_ACC', 'PHIC_PACKAGES', 'PHIC_MCP', \n",
    "            'PHIC_PCB1', 'PHIC_MALARIA', 'PHIC_TBDOTS', 'PHIC_ABTC', \n",
    "            'PHIC_NCP', 'PHIC_OTH', 'REF1_CODE', 'REF1_DIST', 'REF1_TRAVEL',\n",
    "            'REF2_CODE', 'REF2_DIST', 'REF2_TRAVEL',\n",
    "            'REF3_CODE', 'REF3_DIST', 'REF3_TRAVEL']\n",
    "\n",
    "rhu = rhu[cols]\n",
    "\n",
    "# Using the health facility list, complete the RHU data\n",
    "rhu.loc[rhu['REF1_CODE']==0, 'REF_CODE'] = rhu[rhu['REF1_CODE']==0]['REF1_NAME'].map(HF_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:55.095296Z",
     "start_time": "2020-08-10T18:34:55.084326Z"
    }
   },
   "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>SHORT_HFCODE</th>\n",
       "      <th>REF_CODE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2228</td>\n",
       "      <td>{3313}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>6698</td>\n",
       "      <td>{3313}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>147</td>\n",
       "      <td>{2703}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>2033</td>\n",
       "      <td>{3667}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>2184</td>\n",
       "      <td>{2703}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    SHORT_HFCODE REF_CODE\n",
       "8           2228   {3313}\n",
       "20          6698   {3313}\n",
       "29           147   {2703}\n",
       "36          2033   {3667}\n",
       "37          2184   {2703}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "temp = rhu[['SHORT_HFCODE', 'REF_CODE']].dropna().copy()\n",
    "\n",
    "# This dataframe contains the HF codes of one health facility to other facilities.\n",
    "temp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:55.542570Z",
     "start_time": "2020-08-10T18:34:55.504672Z"
    }
   },
   "outputs": [],
   "source": [
    "# Out of all the mapped referred facilities, the closest facility is used as the actual referred facility\n",
    "# Slight pre-processing: \"place\" the health facilities without coordinates to southwest of southwest Philippine boundary (lat_min - 10, long_min - 20) or northeast of northeast Philippine boundary\n",
    "temp_dict = pd.DataFrame(temp.apply(lambda x: min([(vincenty((latlong_dict['LAT'][x['SHORT_HFCODE']] if latlong_dict['LAT'][x['SHORT_HFCODE']]==latlong_dict['LAT'][x['SHORT_HFCODE']] else lat_min-10,\n",
    "                                                  latlong_dict['LONG'][x['SHORT_HFCODE']] if latlong_dict['LONG'][x['SHORT_HFCODE']]==latlong_dict['LONG'][x['SHORT_HFCODE']] else long_min-20), \n",
    "                                                 (latlong_dict['LAT'][i] if latlong_dict['LAT'][i]==latlong_dict['LAT'][i] else lat_max+10,\n",
    "                                                  latlong_dict['LONG'][i] if latlong_dict['LONG'][i]==latlong_dict['LONG'][i] else long_max+20)).km, i, x['SHORT_HFCODE']) for i in x['REF_CODE']], key=lambda x: x[0]), axis=1).tolist()).set_index(2).to_dict()\n",
    "\n",
    "rhu['REF_CODE'] = rhu['SHORT_HFCODE'].map(temp_dict[1])\n",
    "rhu.loc[rhu['REF1_CODE']!=0, 'REF_CODE'] = rhu.loc[rhu['REF1_CODE']!=0, 'REF1_CODE']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Fill missing REF2 Codes\n",
    "\n",
    "The data contains up to three referred facilities so the same processing is performed for the second and third HF code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:56.768895Z",
     "start_time": "2020-08-10T18:34:56.716037Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu2.loc[rhu2['REF2_CODE']==0, 'REF_CODE'] = rhu2[rhu2['REF2_CODE']==0]['REF2_NAME'].map(HF_dict)\n",
    "\n",
    "temp = rhu2[['SHORT_HFCODE', 'REF_CODE']].dropna().copy()\n",
    "temp_dict = pd.DataFrame(temp.apply(lambda x: min([(vincenty((latlong_dict['LAT'][x['SHORT_HFCODE']] if latlong_dict['LAT'][x['SHORT_HFCODE']]==latlong_dict['LAT'][x['SHORT_HFCODE']] else lat_min-10,\n",
    "                                                  latlong_dict['LONG'][x['SHORT_HFCODE']] if latlong_dict['LONG'][x['SHORT_HFCODE']]==latlong_dict['LONG'][x['SHORT_HFCODE']] else long_min-20), \n",
    "                                                 (latlong_dict['LAT'][i] if latlong_dict['LAT'][i]==latlong_dict['LAT'][i] else lat_max+10,\n",
    "                                                  latlong_dict['LONG'][i] if latlong_dict['LONG'][i]==latlong_dict['LONG'][i] else long_max+20)).km, i, x['SHORT_HFCODE']) for i in x['REF_CODE']], key=lambda x: x[0]), axis=1).tolist()).set_index(2).to_dict()\n",
    "\n",
    "rhu2['REF_CODE'] = rhu2['SHORT_HFCODE'].map(temp_dict[1])\n",
    "rhu2.loc[rhu2['REF2_CODE']!=0, 'REF_CODE'] = rhu2.loc[rhu2['REF2_CODE']!=0, 'REF2_CODE']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Fill missing REF3 Codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:58.843245Z",
     "start_time": "2020-08-10T18:34:58.762299Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu3.loc[rhu3['REF3_CODE']==0, 'REF_CODE'] = rhu3[rhu3['REF3_CODE']==0]['REF3_NAME'].map(HF_dict)\n",
    "\n",
    "temp = rhu3[['SHORT_HFCODE', 'REF_CODE']].dropna().copy()\n",
    "temp_dict = pd.DataFrame(temp.apply(lambda x: min([(vincenty((latlong_dict['LAT'][x['SHORT_HFCODE']] if latlong_dict['LAT'][x['SHORT_HFCODE']]==latlong_dict['LAT'][x['SHORT_HFCODE']] else lat_min-10,\n",
    "                                                  latlong_dict['LONG'][x['SHORT_HFCODE']] if latlong_dict['LONG'][x['SHORT_HFCODE']]==latlong_dict['LONG'][x['SHORT_HFCODE']] else long_min-20), \n",
    "                                                 (latlong_dict['LAT'][i] if latlong_dict['LAT'][i]==latlong_dict['LAT'][i] else lat_max+10,\n",
    "                                                  latlong_dict['LONG'][i] if latlong_dict['LONG'][i]==latlong_dict['LONG'][i] else long_max+20)).km, i, x['SHORT_HFCODE']) for i in x['REF_CODE']], key=lambda x: x[0]), axis=1).tolist()).set_index(2).to_dict()\n",
    "\n",
    "rhu3['REF_CODE'] = rhu3['SHORT_HFCODE'].map(temp_dict[1])\n",
    "rhu3.loc[rhu3['REF3_CODE']!=0, 'REF_CODE'] = rhu3.loc[rhu3['REF3_CODE']!=0, 'REF3_CODE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:34:59.080050Z",
     "start_time": "2020-08-10T18:34:59.068083Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu.dropna(subset=['REF_CODE'], inplace=True)\n",
    "rhu2.dropna(subset=['REF_CODE'], inplace=True)\n",
    "rhu3.dropna(subset=['REF_CODE'], inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-03-01T08:03:19.170179Z",
     "start_time": "2020-03-01T08:03:19.165224Z"
    }
   },
   "source": [
    "##### Combine the processed dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:35:00.127287Z",
     "start_time": "2020-08-10T18:35:00.118312Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu.rename({'REF1_DIST':'REF_DIST', 'REF1_TRAVEL':'REF_TRAVEL', 'REF1_NAME':'REF_NAME'}, axis=1, inplace=True)\n",
    "rhu2.rename({'REF2_DIST':'REF_DIST', 'REF2_TRAVEL':'REF_TRAVEL', 'REF1_NAME':'REF_NAME'}, axis=1, inplace=True)\n",
    "rhu3.rename({'REF3_DIST':'REF_DIST', 'REF3_TRAVEL':'REF_TRAVEL', 'REF1_NAME':'REF_NAME'}, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:35:00.691341Z",
     "start_time": "2020-08-10T18:35:00.685359Z"
    }
   },
   "outputs": [],
   "source": [
    "cols2 = ['SHORT_HFCODE', 'REF_CODE']\n",
    "rhu_edges = rhu[cols2].append(rhu2[cols2]).append(rhu3[cols2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:35:00.952090Z",
     "start_time": "2020-08-10T18:35:00.948093Z"
    }
   },
   "outputs": [],
   "source": [
    "# Add a column identifying the type of facility for the later network analysis\n",
    "\n",
    "rhu_edges['HF_TYPE'] = 'RHU'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:35:01.202968Z",
     "start_time": "2020-08-10T18:35:01.183018Z"
    }
   },
   "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>SHORT_HFCODE</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LONG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>29,424.00</td>\n",
       "      <td>25,752.00</td>\n",
       "      <td>25,752.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>19,505.19</td>\n",
       "      <td>12.27</td>\n",
       "      <td>122.49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>11,203.47</td>\n",
       "      <td>7.11</td>\n",
       "      <td>4.42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>9,724.75</td>\n",
       "      <td>9.52</td>\n",
       "      <td>121.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>19,799.50</td>\n",
       "      <td>12.73</td>\n",
       "      <td>122.38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>29,542.25</td>\n",
       "      <td>14.81</td>\n",
       "      <td>124.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>38,002.00</td>\n",
       "      <td>1,000.00</td>\n",
       "      <td>126.58</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       SHORT_HFCODE       LAT      LONG\n",
       "count     29,424.00 25,752.00 25,752.00\n",
       "mean      19,505.19     12.27    122.49\n",
       "std       11,203.47      7.11      4.42\n",
       "min            1.00      0.00      0.00\n",
       "25%        9,724.75      9.52    121.02\n",
       "50%       19,799.50     12.73    122.38\n",
       "75%       29,542.25     14.81    124.16\n",
       "max       38,002.00  1,000.00    126.58"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "HF_list[['SHORT_HFCODE', 'LAT', 'LONG']].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Map lat-long to the HF codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:38.861865Z",
     "start_time": "2020-08-10T18:46:38.748169Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges['source_lat'] = rhu_edges['SHORT_HFCODE'].map(latlong_dict['LAT'])\n",
    "rhu_edges['source_long'] = rhu_edges['SHORT_HFCODE'].map(latlong_dict['LONG'])\n",
    "rhu_edges['target_lat'] = rhu_edges['REF_CODE'].map(latlong_dict['LAT'])\n",
    "rhu_edges['target_long'] = rhu_edges['REF_CODE'].map(latlong_dict['LONG'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Set lat-long outside PH to NaN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:40.195569Z",
     "start_time": "2020-08-10T18:46:40.180642Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges.loc[~((rhu_edges['source_lat'].between(lat_min, lat_max)) & (rhu_edges['source_long'].between(\n",
    "    long_min, long_max))), ['source_lat', 'source_long']] = np.nan\n",
    "rhu_edges.loc[~((rhu_edges['target_lat'].between(lat_min, lat_max)) & (rhu_edges['target_long'].between(\n",
    "    long_min, long_max))), ['target_lat', 'target_long']] = np.nan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Measure distance using lat-long for non-NaN HF pairs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:40.935464Z",
     "start_time": "2020-08-10T18:46:40.596159Z"
    }
   },
   "outputs": [],
   "source": [
    "missing_latlong = ~rhu_edges[['source_lat', 'source_long', 'target_lat', 'target_long']].isnull().sum(axis=1).astype(bool)\n",
    "rhu_edges.loc[missing_latlong, 'DIST'] = rhu_edges.loc[missing_latlong, ['source_lat', 'source_long', 'target_lat', 'target_long']].apply(lambda x: vincenty((x['source_lat'], x['source_long']), (x['target_lat'], x['target_long'])).km, axis=1)\n",
    "\n",
    "rhu_edges.loc[rhu_edges['DIST']==0, 'DIST'] = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Check for outliers\n",
    "In this case, referral between facilities that are too high are classified as outliers. The threshold to be used for filtering out these outliers is subject to the researcher's choice. The unit of distance is in kilometers. Note that since the distance is extracted from coordinates, errors in encoding will change the actual coordinates of the facilities, therefore making changes to the distance (e.g. dividing by a constant factor) cannot be done. Instead, other imputing techinques are used to fill for missing distance data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:41.979471Z",
     "start_time": "2020-08-10T18:46:41.481073Z"
    }
   },
   "outputs": [
    {
     "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": [
    "B = rhu_edges.boxplot('DIST', return_type='both')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:42.584794Z",
     "start_time": "2020-08-10T18:46:42.578796Z"
    }
   },
   "outputs": [],
   "source": [
    "outliers = [i.get_ydata()[1] for i in B.lines['whiskers']]\n",
    "rhu_edges.loc[rhu_edges['DIST'] > outliers[1], 'DIST'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:52.361936Z",
     "start_time": "2020-08-10T18:46:52.357947Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.0006040641840882818, 88.56724901828711]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:52.732984Z",
     "start_time": "2020-08-10T18:46:52.714035Z"
    }
   },
   "outputs": [],
   "source": [
    "muni_dict = HF_list[['SHORT_HFCODE', 'MUNI_CITY', 'PROVINCE']].set_index('SHORT_HFCODE').to_dict()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Impute distance from municipality\n",
    "With the municipality information of the facilities, the distance of different health units are imputed using the median. The assumption is that within the same municipality, the distance of referring facilities are more or less similar. For tracking, the source of the distance information is stored."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:53.374218Z",
     "start_time": "2020-08-10T18:46:53.343302Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges['muni_city'] = rhu_edges['SHORT_HFCODE'].map(muni_dict['MUNI_CITY'])\n",
    "mean_dist_city = rhu_edges.groupby('muni_city')['DIST'].median().to_dict()\n",
    "\n",
    "imputed_muni = ~(rhu_edges.loc[rhu_edges['DIST'].isnull(), 'muni_city'].map(mean_dist_city).isnull())\n",
    "imputed_muni = imputed_muni[imputed_muni].index\n",
    "rhu_edges.loc[imputed_muni, \"IMPUTED\"] = \"MUNI\" \n",
    "rhu_edges.loc[rhu_edges['DIST'].isnull(), 'DIST'] = rhu_edges.loc[rhu_edges['DIST'].isnull(), 'muni_city'].map(mean_dist_city)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Impute distance from province\n",
    "For those facilities without municipality information, the province is used."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:53.828715Z",
     "start_time": "2020-08-10T18:46:53.772865Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges['province'] = rhu_edges['SHORT_HFCODE'].map(muni_dict['PROVINCE'])\n",
    "mean_dist_prov = rhu_edges.groupby('province')['DIST'].median().to_dict()\n",
    "\n",
    "imputed_prov = ~(rhu_edges.loc[rhu_edges['DIST'].isnull(), 'province'].map(mean_dist_prov).isnull())\n",
    "imputed_prov = imputed_prov[imputed_prov].index\n",
    "\n",
    "rhu_edges.loc[imputed_prov, \"IMPUTED\"] = \"PROV\"\n",
    "rhu_edges.loc[rhu_edges['DIST'].isnull(), 'DIST'] = rhu_edges.loc[rhu_edges['DIST'].isnull(), 'province'].map(mean_dist_prov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:53.974282Z",
     "start_time": "2020-08-10T18:46:53.969295Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "17"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rhu_edges['DIST'].isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If after all these, the referring facilities still do not have distance information, the connection between the facilities is dropped. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:54.278528Z",
     "start_time": "2020-08-10T18:46:54.273494Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges.dropna(subset=['DIST'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:54.494949Z",
     "start_time": "2020-08-10T18:46:54.470967Z"
    }
   },
   "outputs": [],
   "source": [
    "prov_HF_dict = rhu_edges.groupby('province')[['SHORT_HFCODE', 'REF_CODE']].agg(set).rename({'SHORT_HFCODE':'RHU', 'REF_CODE':'HOSP'}, axis=1).to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:54.646506Z",
     "start_time": "2020-08-10T18:46:54.639548Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges['REF_CODES'] = [set(rhu_edges[~rhu_edges[['target_lat', 'target_long']].isnull().sum(axis=1).astype(bool)]['REF_CODE'])] * len(rhu_edges)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:46:54.862996Z",
     "start_time": "2020-08-10T18:46:54.858043Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges = rhu_edges[rhu_edges['REF_CODE']!=rhu_edges['SHORT_HFCODE']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Connect nearest neighbors\n",
    "The referrals above are based on actual data, i.e. actual referrals from facility to facility. This list is supplemented with the three nearest facilities, regardless of being the same facilities as the actual data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:49.846281Z",
     "start_time": "2020-08-10T18:46:55.331148Z"
    }
   },
   "outputs": [],
   "source": [
    "n = 3 #num of nearest neighbors to connect\n",
    "temp = rhu_edges[['SHORT_HFCODE', 'REF_CODES']].dropna().drop_duplicates(subset='SHORT_HFCODE').copy()\n",
    "df_neighbors = pd.DataFrame(temp.apply(lambda x: sorted([(vincenty((latlong_dict['LAT'][x['SHORT_HFCODE']] if latlong_dict['LAT'][x['SHORT_HFCODE']]==latlong_dict['LAT'][x['SHORT_HFCODE']] else lat_min-10,\n",
    "                                                  latlong_dict['LONG'][x['SHORT_HFCODE']] if latlong_dict['LONG'][x['SHORT_HFCODE']]==latlong_dict['LONG'][x['SHORT_HFCODE']] else long_min-20), \n",
    "                                                 (latlong_dict['LAT'][i] if latlong_dict['LAT'][i]==latlong_dict['LAT'][i] else lat_max+10,\n",
    "                                                  latlong_dict['LONG'][i] if latlong_dict['LONG'][i]==latlong_dict['LONG'][i] else long_max+20)).km, i, x['SHORT_HFCODE']) for i in x['REF_CODES'] if i!=x['SHORT_HFCODE']], key=lambda x: x[0])[:n], axis=1).tolist())#.set_index(2).to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:49.993922Z",
     "start_time": "2020-08-10T18:48:49.986905Z"
    }
   },
   "outputs": [],
   "source": [
    "df_neighbors_edges = pd.DataFrame(df_neighbors[0].append(df_neighbors[1]).append(df_neighbors[2]).tolist(), columns=['DIST', 'REF_CODE', 'SHORT_HFCODE'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:50.127563Z",
     "start_time": "2020-08-10T18:48:50.124538Z"
    }
   },
   "outputs": [],
   "source": [
    "df_neighbors_edges['IMPUTED'] = 'NEAREST_NEIGHBOR'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:50.263167Z",
     "start_time": "2020-08-10T18:48:50.253249Z"
    }
   },
   "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>SHORT_HFCODE</th>\n",
       "      <th>REF_CODE</th>\n",
       "      <th>DIST</th>\n",
       "      <th>IMPUTED</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>25</td>\n",
       "      <td>32,170.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>105</td>\n",
       "      <td>5,940.00</td>\n",
       "      <td>0.41</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>106</td>\n",
       "      <td>3,313.00</td>\n",
       "      <td>11.13</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>137</td>\n",
       "      <td>3,313.00</td>\n",
       "      <td>11.29</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1760</td>\n",
       "      <td>6,513.00</td>\n",
       "      <td>5.98</td>\n",
       "      <td>NaN</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>2296</th>\n",
       "      <td>6814</td>\n",
       "      <td>273.00</td>\n",
       "      <td>51.12</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2297</th>\n",
       "      <td>7045</td>\n",
       "      <td>273.00</td>\n",
       "      <td>15.07</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2298</th>\n",
       "      <td>7696</td>\n",
       "      <td>273.00</td>\n",
       "      <td>29.09</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2299</th>\n",
       "      <td>8861</td>\n",
       "      <td>273.00</td>\n",
       "      <td>16.35</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2300</th>\n",
       "      <td>27898</td>\n",
       "      <td>237.00</td>\n",
       "      <td>41.68</td>\n",
       "      <td>MUNI</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5925 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      SHORT_HFCODE  REF_CODE  DIST IMPUTED\n",
       "0               25 32,170.00  0.00     NaN\n",
       "1              105  5,940.00  0.41     NaN\n",
       "2              106  3,313.00 11.13     NaN\n",
       "4              137  3,313.00 11.29     NaN\n",
       "5             1760  6,513.00  5.98     NaN\n",
       "...            ...       ...   ...     ...\n",
       "2296          6814    273.00 51.12     NaN\n",
       "2297          7045    273.00 15.07     NaN\n",
       "2298          7696    273.00 29.09     NaN\n",
       "2299          8861    273.00 16.35     NaN\n",
       "2300         27898    237.00 41.68    MUNI\n",
       "\n",
       "[5925 rows x 4 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rhu_edges[['SHORT_HFCODE', 'REF_CODE', 'DIST', 'IMPUTED']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:50.398837Z",
     "start_time": "2020-08-10T18:48:50.391855Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges = rhu_edges[['SHORT_HFCODE', 'REF_CODE', 'DIST', 'IMPUTED']].append(df_neighbors_edges)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:50.527493Z",
     "start_time": "2020-08-10T18:48:50.520497Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_edges.loc[rhu_edges['DIST'] > outliers[1], 'DIST'] = np.nan\n",
    "rhu_edges.loc[rhu_edges['DIST']==0, 'DIST'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:50.662103Z",
     "start_time": "2020-08-10T18:48:50.648138Z"
    }
   },
   "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>DIST</th>\n",
       "      <th>REF_CODE</th>\n",
       "      <th>SHORT_HFCODE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>12,157.00</td>\n",
       "      <td>12,330.00</td>\n",
       "      <td>12,330.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>16.11</td>\n",
       "      <td>4,600.62</td>\n",
       "      <td>5,816.23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>16.55</td>\n",
       "      <td>7,159.54</td>\n",
       "      <td>6,469.21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>10.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.39</td>\n",
       "      <td>622.00</td>\n",
       "      <td>2,307.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>11.07</td>\n",
       "      <td>2,850.00</td>\n",
       "      <td>4,234.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>22.87</td>\n",
       "      <td>5,139.00</td>\n",
       "      <td>7,062.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>88.57</td>\n",
       "      <td>261,101.00</td>\n",
       "      <td>36,628.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           DIST   REF_CODE  SHORT_HFCODE\n",
       "count 12,157.00  12,330.00     12,330.00\n",
       "mean      16.11   4,600.62      5,816.23\n",
       "std       16.55   7,159.54      6,469.21\n",
       "min        0.00       1.00         10.00\n",
       "25%        3.39     622.00      2,307.00\n",
       "50%       11.07   2,850.00      4,234.00\n",
       "75%       22.87   5,139.00      7,062.00\n",
       "max       88.57 261,101.00     36,628.00"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rhu_edges.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### BHS\n",
    "The same processing done for the rural health units is performed for the barangay health stations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:48:58.675223Z",
     "start_time": "2020-08-10T18:48:50.784773Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs = pd.read_excel('bhs2018.xlsx', sheet_name='MAIN', na_values='None')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:51:12.857549Z",
     "start_time": "2020-08-10T18:51:02.056784Z"
    }
   },
   "outputs": [],
   "source": [
    "str_cols = ['HF_NAME', 'REGION', 'PROVINCE', 'MUNI_CITY', 'BRGY',\n",
    "            'STREET_NAME', 'BUILDING', 'FACILITY_HEAD', 'DETACHED', \n",
    "            'BRGYS', 'RHU_NAME', 'RHU_SAME_CTY', 'RHU_NOTSAME_CTY',\n",
    "            'AMB_ACCESS']\n",
    "code_cols = ['id', 'RHU_CODE']\n",
    "float_cols = ['RHU_DIST', 'RHU_TRAVEL']\n",
    "# int_cols = ['CATCHMENT', 'MD_NO', 'MD_AUG', 'MD_TOTAL',\n",
    "#        'MD_FT', 'MD_PT', 'MD_VISIT', 'RN_NO', 'RN_AUG', 'RN_TOTAL', 'RN_FT',\n",
    "#        'RN_PT', 'RN_VISIT', 'MW_NO', 'MW_AUG', 'MW_TOTAL', 'MW_FT', 'MW_PT',\n",
    "#        'MW_VISIT', 'BHW_NO']\n",
    "\n",
    "bhs[str_cols] = bhs[str_cols].apply(lambda x: x.str.upper().str.strip())\n",
    "bhs[code_cols] = bhs[code_cols].fillna(0).astype(int)\n",
    "bhs[float_cols] = bhs[float_cols].astype(float)\n",
    "\n",
    "bhs[str_cols] = bhs[str_cols].fillna('UNKNOWN')\n",
    "bhs['SHORT_HFCODE'] = bhs['HF_CODE'].apply(lambda x: int(x[-6:]))\n",
    "bhs.to_excel('cleaned/bhs_cleaned.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Fill missing RHU Codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:51:20.683017Z",
     "start_time": "2020-08-10T18:51:13.803276Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs = pd.read_excel('cleaned/bhs_cleaned.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T18:51:20.925293Z",
     "start_time": "2020-08-10T18:51:20.827516Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs.loc[bhs['RHU_CODE']==0, 'REF_CODE'] = bhs[bhs['RHU_CODE']==0]['RHU_NAME'].map(HF_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:20.789476Z",
     "start_time": "2020-08-10T19:02:20.724633Z"
    }
   },
   "outputs": [],
   "source": [
    "temp = bhs[['SHORT_HFCODE', 'REF_CODE']].dropna().copy()\n",
    "temp_dict = pd.DataFrame(temp.apply(lambda x: min([(vincenty((latlong_dict['LAT'][x['SHORT_HFCODE']] if latlong_dict['LAT'][x['SHORT_HFCODE']]==latlong_dict['LAT'][x['SHORT_HFCODE']] else lat_min-10,\n",
    "                                                  latlong_dict['LONG'][x['SHORT_HFCODE']] if latlong_dict['LONG'][x['SHORT_HFCODE']]==latlong_dict['LONG'][x['SHORT_HFCODE']] else long_min-20), \n",
    "                                                 (latlong_dict['LAT'][i] if latlong_dict['LAT'][i]==latlong_dict['LAT'][i] else lat_max+10,\n",
    "                                                  latlong_dict['LONG'][i] if latlong_dict['LONG'][i]==latlong_dict['LONG'][i] else long_max+20)).km, i, x['SHORT_HFCODE']) for i in x['REF_CODE']], key=lambda x: x[0]), axis=1).tolist()).set_index(2).to_dict()\n",
    "\n",
    "bhs['REF_CODE'] = bhs['SHORT_HFCODE'].map(temp_dict[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:21.865066Z",
     "start_time": "2020-08-10T19:02:21.859115Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs.dropna(subset=['REF_CODE'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:22.334020Z",
     "start_time": "2020-08-10T19:02:22.331028Z"
    }
   },
   "outputs": [],
   "source": [
    "cols = ['SHORT_HFCODE', 'REF_CODE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:22.680157Z",
     "start_time": "2020-08-10T19:02:22.675170Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs = bhs[cols]\n",
    "bhs['HF_TYPE'] = 'BHS'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:23.379285Z",
     "start_time": "2020-08-10T19:02:23.227554Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs['source_lat'] = bhs['SHORT_HFCODE'].map(latlong_dict['LAT'])\n",
    "bhs['source_long'] = bhs['SHORT_HFCODE'].map(latlong_dict['LONG'])\n",
    "bhs['target_lat'] = bhs['REF_CODE'].map(latlong_dict['LAT'])\n",
    "bhs['target_long'] = bhs['REF_CODE'].map(latlong_dict['LONG'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:24.574132Z",
     "start_time": "2020-08-10T19:02:24.557179Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs.loc[~((bhs['source_lat'].between(lat_min, lat_max)) & (bhs['source_long'].between(\n",
    "    long_min, long_max))), ['source_lat', 'source_long']] = np.nan\n",
    "bhs.loc[~((bhs['target_lat'].between(lat_min, lat_max)) & (bhs['target_long'].between(\n",
    "    long_min, long_max))), ['target_lat', 'target_long']] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:25.823250Z",
     "start_time": "2020-08-10T19:02:25.784322Z"
    }
   },
   "outputs": [],
   "source": [
    "missing_latlong = ~bhs[['source_lat', 'source_long', 'target_lat', 'target_long']].isnull().sum(axis=1).astype(bool)\n",
    "bhs.loc[missing_latlong, 'DIST'] = bhs.loc[missing_latlong, ['source_lat', 'source_long', 'target_lat', 'target_long']].apply(lambda x: vincenty((x['source_lat'], x['source_long']), (x['target_lat'], x['target_long'])).km, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:26.913271Z",
     "start_time": "2020-08-10T19:02:26.905289Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs.loc[bhs['DIST']==0, 'DIST'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:27.444692Z",
     "start_time": "2020-08-10T19:02:27.355434Z"
    }
   },
   "outputs": [
    {
     "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": [
    "B = bhs.boxplot('DIST', return_type='both')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:27.764566Z",
     "start_time": "2020-08-10T19:02:27.760577Z"
    }
   },
   "outputs": [],
   "source": [
    "outliers = [i.get_ydata()[1] for i in B.lines['whiskers']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:28.264561Z",
     "start_time": "2020-08-10T19:02:28.260572Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.0024470114519412256, 16.988283224538524]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:28.624570Z",
     "start_time": "2020-08-10T19:02:28.616592Z"
    }
   },
   "outputs": [],
   "source": [
    "outliers = [i.get_ydata()[1] for i in B.lines['whiskers']]\n",
    "bhs.loc[bhs['DIST'] > outliers[1], 'DIST'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:28.917558Z",
     "start_time": "2020-08-10T19:02:28.898608Z"
    }
   },
   "outputs": [],
   "source": [
    "muni_dict = HF_list[['SHORT_HFCODE', 'MUNI_CITY', 'PROVINCE']].set_index('SHORT_HFCODE').to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:31.109115Z",
     "start_time": "2020-08-10T19:02:31.075036Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs['muni_city'] = bhs['SHORT_HFCODE'].map(muni_dict['MUNI_CITY'])\n",
    "mean_dist_city = bhs.groupby('muni_city')['DIST'].mean().to_dict()\n",
    "\n",
    "imputed_muni = ~(bhs.loc[bhs['DIST'].isnull(), 'muni_city'].map(mean_dist_city).isnull())\n",
    "imputed_muni = imputed_muni[imputed_muni].index\n",
    "bhs.loc[imputed_muni, \"IMPUTED\"] = \"MUNI\" \n",
    "bhs.loc[bhs['DIST'].isnull(), 'DIST'] = bhs.loc[bhs['DIST'].isnull(), 'muni_city'].map(mean_dist_city)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:31.330394Z",
     "start_time": "2020-08-10T19:02:31.307416Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs['province'] = bhs['SHORT_HFCODE'].map(muni_dict['PROVINCE'])\n",
    "mean_dist_prov = bhs.groupby('province')['DIST'].mean().to_dict()\n",
    "\n",
    "imputed_prov = ~(bhs.loc[bhs['DIST'].isnull(), 'province'].map(mean_dist_prov).isnull())\n",
    "imputed_prov = imputed_prov[imputed_prov].index\n",
    "\n",
    "bhs.loc[imputed_prov, \"IMPUTED\"] = \"PROV\"\n",
    "bhs.loc[bhs['DIST'].isnull(), 'DIST'] = bhs.loc[bhs['DIST'].isnull(), 'province'].map(mean_dist_prov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:31.536804Z",
     "start_time": "2020-08-10T19:02:31.531817Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bhs['DIST'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:31.742252Z",
     "start_time": "2020-08-10T19:02:31.737267Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs.dropna(subset=['DIST'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:31.948701Z",
     "start_time": "2020-08-10T19:02:31.943715Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs = bhs[['SHORT_HFCODE', 'REF_CODE', 'DIST', 'IMPUTED']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:32.134205Z",
     "start_time": "2020-08-10T19:02:32.130217Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs = bhs[bhs['REF_CODE'] != bhs['SHORT_HFCODE']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:32.330716Z",
     "start_time": "2020-08-10T19:02:32.317717Z"
    }
   },
   "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>DIST</th>\n",
       "      <th>IMPUTED</th>\n",
       "      <th>REF_CODE</th>\n",
       "      <th>SHORT_HFCODE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>32,170.00</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.41</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5,940.00</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11.13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3,313.00</td>\n",
       "      <td>106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>11.29</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3,313.00</td>\n",
       "      <td>137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.98</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6,513.00</td>\n",
       "      <td>1760</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>19297</th>\n",
       "      <td>8.88</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7,207.00</td>\n",
       "      <td>17805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19325</th>\n",
       "      <td>3.56</td>\n",
       "      <td>NaN</td>\n",
       "      <td>79.00</td>\n",
       "      <td>17893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19345</th>\n",
       "      <td>0.65</td>\n",
       "      <td>NaN</td>\n",
       "      <td>27,529.00</td>\n",
       "      <td>27530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19367</th>\n",
       "      <td>1.68</td>\n",
       "      <td>NaN</td>\n",
       "      <td>169.00</td>\n",
       "      <td>29236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19368</th>\n",
       "      <td>12.67</td>\n",
       "      <td>NaN</td>\n",
       "      <td>169.00</td>\n",
       "      <td>29238</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12793 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       DIST IMPUTED  REF_CODE  SHORT_HFCODE\n",
       "0      0.00     NaN 32,170.00            25\n",
       "1      0.41     NaN  5,940.00           105\n",
       "2     11.13     NaN  3,313.00           106\n",
       "4     11.29     NaN  3,313.00           137\n",
       "5      5.98     NaN  6,513.00          1760\n",
       "...     ...     ...       ...           ...\n",
       "19297  8.88     NaN  7,207.00         17805\n",
       "19325  3.56     NaN     79.00         17893\n",
       "19345  0.65     NaN 27,529.00         27530\n",
       "19367  1.68     NaN    169.00         29236\n",
       "19368 12.67     NaN    169.00         29238\n",
       "\n",
       "[12793 rows x 4 columns]"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "edge_list = rhu_edges.append(bhs)\n",
    "edge_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:35.098781Z",
     "start_time": "2020-08-10T19:02:35.077802Z"
    }
   },
   "outputs": [],
   "source": [
    "edge_list = edge_list.groupby(['REF_CODE', 'SHORT_HFCODE'])[['IMPUTED', 'DIST']].agg({'DIST':'mean', 'IMPUTED':'first'}).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:39.737440Z",
     "start_time": "2020-08-10T19:02:39.182223Z"
    }
   },
   "outputs": [],
   "source": [
    "edge_list.to_excel('edge_list.xlsx', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Network Analysis\n",
    "\n",
    "Using the facilities as nodes and the referral as edges, a tripartite network of BHS, RHU, and hospitals is created. The characteristics of the healthcare provider network for different regions are characterized and explored in this section. The main metric used in the analysis is degree and path length. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:02:40.940396Z",
     "start_time": "2020-08-10T19:02:40.935410Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import networkx as nx\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import itertools\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:55.648453Z",
     "start_time": "2020-08-10T19:19:52.428885Z"
    }
   },
   "outputs": [],
   "source": [
    "nodes = pd.read_excel(\"nodeslist.xlsx\")\n",
    "edges = pd.read_excel(\"edge_list.xlsx\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the succeeding analysis, the imputed distances using the nearest neighbors are not included."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:57.945781Z",
     "start_time": "2020-08-10T19:19:57.939797Z"
    }
   },
   "outputs": [],
   "source": [
    "edges = edges[(edges['IMPUTED']!='NEAREST_NEIGHBOR') | (edges['SHORT_HFCODE']!=6108)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:58.293853Z",
     "start_time": "2020-08-10T19:19:58.287834Z"
    }
   },
   "outputs": [],
   "source": [
    "nodes.drop(['LAT', 'LONG'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:58.733768Z",
     "start_time": "2020-08-10T19:19:58.637900Z"
    }
   },
   "outputs": [],
   "source": [
    "G = nx.from_pandas_edgelist(edges, source='SHORT_HFCODE', target='REF_CODE', edge_attr='DIST', create_using=nx.DiGraph)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:59.027972Z",
     "start_time": "2020-08-10T19:19:59.024990Z"
    }
   },
   "outputs": [],
   "source": [
    "nodes.index = nodes['SHORT_HFCODE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:59.379043Z",
     "start_time": "2020-08-10T19:19:59.315180Z"
    }
   },
   "outputs": [],
   "source": [
    "node_attrib = nodes.to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:19:59.881219Z",
     "start_time": "2020-08-10T19:19:59.668745Z"
    }
   },
   "outputs": [],
   "source": [
    "for col in nodes.columns:\n",
    "    nx.set_node_attributes(G, node_attrib[col], col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:00.193344Z",
     "start_time": "2020-08-10T19:20:00.171404Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs_nodes = [i for i, j in G.nodes(data=True) if j.get('HF_TYPE')=='BARANGAY HEALTH STATION']\n",
    "rhu_nodes = [i for i, j in G.nodes(data=True) if j.get('HF_TYPE')=='RURAL HEALTH UNIT']\n",
    "hosp_nodes = [i for i, j in G.nodes(data=True) if ((j.get('HF_TYPE')=='HOSPITAL') or (j.get('HF_TYPE')=='INFIRMARY'))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Degree (Hospital)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:00.485588Z",
     "start_time": "2020-08-10T19:20:00.482602Z"
    }
   },
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:00.867576Z",
     "start_time": "2020-08-10T19:20:00.787754Z"
    }
   },
   "outputs": [],
   "source": [
    "region_deg = {}\n",
    "for region in nodes['REGION'].unique():\n",
    "    deg = list(dict(G.in_degree(nodes[((nodes['HF_TYPE']=='INFIRMARY') | (nodes['HF_TYPE']=='HOSPITAL')) & (nodes['REGION']==region)]['SHORT_HFCODE'])).values())\n",
    "    region_deg[region] = deg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:01.160765Z",
     "start_time": "2020-08-10T19:20:01.154773Z"
    }
   },
   "outputs": [],
   "source": [
    "df_deg = pd.DataFrame(list(itertools.chain(*[list(zip(len(j) * [i], j)) for i, j in region_deg.items()]))).rename({0:'Region', 1:'Degree'}, axis=1)#.boxplot(by=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:01.459990Z",
     "start_time": "2020-08-10T19:20:01.456000Z"
    }
   },
   "outputs": [],
   "source": [
    "region_map = {'AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)':'ARMM',\n",
    "       'CORDILLERA ADMINISTRA TIVE REGION (CAR)':'CAR',\n",
    "       'REGION IV-B (MIMAROPA)':'IV-B', 'NATIONAL CAPITAL REGION (NCR)':'NCR',\n",
    "       'REGION X (NORTHERN MINDANAO)':'X', 'REGION XI (DAVAO REGION)':'XI',\n",
    "       'REGION XII (SOCCSKSA RGEN)':'XII', 'REGION XIII (CARAGA)':'XIII',\n",
    "       'REGION I (ILOCOS REGION)':'I', 'REGION II (CAGAYAN VALLEY)':'II',\n",
    "       'REGION III (CENTRAL LUZON)':'III', 'REGION IV-A (CALABAR ZON)':'IV-A',\n",
    "       'REGION V (BICOL REGION)':'V', 'REGION VI (WESTERN VISAYAS)':'VI',\n",
    "       'REGION VII (CENTRAL VISAYAS)':'VII', 'REGION VIII (EASTERN VISAYAS)':'VIII',\n",
    "       'REGION IX (ZAMBOANGA PENINSULA)':'IX'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:01.756165Z",
     "start_time": "2020-08-10T19:20:01.752177Z"
    }
   },
   "outputs": [],
   "source": [
    "df_deg['Region'] = df_deg['Region'].map(region_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:02.059355Z",
     "start_time": "2020-08-10T19:20:02.046390Z"
    }
   },
   "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>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Region</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ARMM</th>\n",
       "      <td>24</td>\n",
       "      <td>6.94</td>\n",
       "      <td>4.50</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>32</td>\n",
       "      <td>9.50</td>\n",
       "      <td>7.00</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>68</td>\n",
       "      <td>14.95</td>\n",
       "      <td>10.00</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>58</td>\n",
       "      <td>9.66</td>\n",
       "      <td>7.00</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>153</td>\n",
       "      <td>17.21</td>\n",
       "      <td>12.50</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>144</td>\n",
       "      <td>11.51</td>\n",
       "      <td>8.00</td>\n",
       "      <td>61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>59</td>\n",
       "      <td>8.32</td>\n",
       "      <td>6.00</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>53</td>\n",
       "      <td>13.17</td>\n",
       "      <td>11.00</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>73</td>\n",
       "      <td>29.92</td>\n",
       "      <td>28.50</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>65</td>\n",
       "      <td>12.10</td>\n",
       "      <td>8.00</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>96</td>\n",
       "      <td>10.84</td>\n",
       "      <td>7.00</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>63</td>\n",
       "      <td>10.23</td>\n",
       "      <td>8.00</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>146</td>\n",
       "      <td>15.92</td>\n",
       "      <td>11.00</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>47</td>\n",
       "      <td>8.58</td>\n",
       "      <td>6.50</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>30</td>\n",
       "      <td>6.85</td>\n",
       "      <td>5.50</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>47</td>\n",
       "      <td>8.10</td>\n",
       "      <td>5.00</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>37</td>\n",
       "      <td>7.58</td>\n",
       "      <td>6.00</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        max  mean  median  count\n",
       "Region                          \n",
       "ARMM     24  6.94    4.50     18\n",
       "CAR      32  9.50    7.00     36\n",
       "I        68 14.95   10.00     39\n",
       "II       58  9.66    7.00     38\n",
       "III     153 17.21   12.50     62\n",
       "IV-A    144 11.51    8.00     61\n",
       "IV-B     59  8.32    6.00     34\n",
       "IX       53 13.17   11.00     23\n",
       "NCR      73 29.92   28.50     48\n",
       "V        65 12.10    8.00     41\n",
       "VI       96 10.84    7.00     50\n",
       "VII      63 10.23    8.00     48\n",
       "VIII    146 15.92   11.00     39\n",
       "X        47  8.58    6.50     36\n",
       "XI       30  6.85    5.50     26\n",
       "XII      47  8.10    5.00     21\n",
       "XIII     37  7.58    6.00     26"
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df_deg.groupby('Region')['Degree']\n",
    " .agg(['min', 'max', 'mean', 'median', 'count']).sort_values(by='Region')\n",
    " [['max', 'mean', 'median', 'count']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Degree Outliers (Hospital)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:02.362576Z",
     "start_time": "2020-08-10T19:20:02.353569Z"
    }
   },
   "outputs": [],
   "source": [
    "upper_fence = df_deg.groupby('Region')['Degree'].quantile(0.75) + 1.5 * (df_deg.groupby('Region')['Degree'].quantile(0.75) - df_deg.groupby('Region')['Degree'].quantile(0.25))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:02.989867Z",
     "start_time": "2020-08-10T19:20:02.652800Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "\n",
    "# fig = plt.figure()\n",
    "sns.boxplot(data=df_deg, x='Region', y='Degree', palette='Blues',\n",
    "            order=upper_fence.sort_values(ascending=False).index)\n",
    "labels = [l.get_text() for l in ax.get_xticklabels()]\n",
    "ax.set_xticklabels(labels, ha='right');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:04.553597Z",
     "start_time": "2020-08-10T19:20:04.540639Z"
    }
   },
   "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>min</th>\n",
       "      <th>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Region</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>ARMM</th>\n",
       "      <td>6.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>3.75</td>\n",
       "      <td>10.75</td>\n",
       "      <td>7.42</td>\n",
       "      <td>7.75</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>1.75</td>\n",
       "      <td>39.75</td>\n",
       "      <td>27.00</td>\n",
       "      <td>33.25</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>0.62</td>\n",
       "      <td>34.62</td>\n",
       "      <td>13.62</td>\n",
       "      <td>5.62</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>4.50</td>\n",
       "      <td>123.50</td>\n",
       "      <td>38.83</td>\n",
       "      <td>27.50</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>8.00</td>\n",
       "      <td>123.00</td>\n",
       "      <td>39.50</td>\n",
       "      <td>13.50</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>2.00</td>\n",
       "      <td>45.00</td>\n",
       "      <td>14.25</td>\n",
       "      <td>5.00</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>16.00</td>\n",
       "      <td>16.00</td>\n",
       "      <td>16.00</td>\n",
       "      <td>16.00</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>4.88</td>\n",
       "      <td>4.88</td>\n",
       "      <td>4.88</td>\n",
       "      <td>4.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>12.00</td>\n",
       "      <td>45.00</td>\n",
       "      <td>31.25</td>\n",
       "      <td>34.00</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>7.62</td>\n",
       "      <td>76.62</td>\n",
       "      <td>27.38</td>\n",
       "      <td>12.62</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>1.62</td>\n",
       "      <td>41.62</td>\n",
       "      <td>14.62</td>\n",
       "      <td>7.62</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>0.25</td>\n",
       "      <td>121.25</td>\n",
       "      <td>34.75</td>\n",
       "      <td>8.75</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>17.00</td>\n",
       "      <td>32.00</td>\n",
       "      <td>24.50</td>\n",
       "      <td>24.50</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>8.00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>9.00</td>\n",
       "      <td>9.00</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>26.50</td>\n",
       "      <td>26.50</td>\n",
       "      <td>26.50</td>\n",
       "      <td>26.50</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>2.62</td>\n",
       "      <td>23.62</td>\n",
       "      <td>13.12</td>\n",
       "      <td>13.12</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         min    max  mean  median  count\n",
       "Region                                  \n",
       "ARMM    6.00   6.00  6.00    6.00      1\n",
       "CAR     3.75  10.75  7.42    7.75      3\n",
       "I       1.75  39.75 27.00   33.25      4\n",
       "II      0.62  34.62 13.62    5.62      3\n",
       "III     4.50 123.50 38.83   27.50      6\n",
       "IV-A    8.00 123.00 39.50   13.50      4\n",
       "IV-B    2.00  45.00 14.25    5.00      4\n",
       "IX     16.00  16.00 16.00   16.00      1\n",
       "NCR     4.88   4.88  4.88    4.88      1\n",
       "V      12.00  45.00 31.25   34.00      4\n",
       "VI      7.62  76.62 27.38   12.62      4\n",
       "VII     1.62  41.62 14.62    7.62      4\n",
       "VIII    0.25 121.25 34.75    8.75      4\n",
       "X      17.00  32.00 24.50   24.50      2\n",
       "XI      8.00  10.00  9.00    9.00      2\n",
       "XII    26.50  26.50 26.50   26.50      1\n",
       "XIII    2.62  23.62 13.12   13.12      2"
      ]
     },
     "execution_count": 183,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_deg['Outlierness'] = df_deg['Degree'] - df_deg['Region'].map(upper_fence)\n",
    "deg_outlier = df_deg[df_deg['Outlierness'] > 0].groupby('Region')['Outlierness'].agg(['min', 'max', 'mean', 'median', 'count'])\n",
    "deg_outlier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Degree (RHU)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:05.681894Z",
     "start_time": "2020-08-10T19:20:05.678868Z"
    }
   },
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:06.101895Z",
     "start_time": "2020-08-10T19:20:06.041899Z"
    }
   },
   "outputs": [],
   "source": [
    "region_deg = {}\n",
    "for region in nodes['REGION'].unique():\n",
    "    deg = list(dict(G.in_degree(nodes[(nodes['HF_TYPE']=='RURAL HEALTH UNIT') & (nodes['REGION']==region)]['SHORT_HFCODE'])).values())\n",
    "    region_deg[region] = deg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:06.404952Z",
     "start_time": "2020-08-10T19:20:06.399942Z"
    }
   },
   "outputs": [],
   "source": [
    "df_deg = pd.DataFrame(list(itertools.chain(*[list(zip(len(j) * [i], j)) for i, j in region_deg.items()]))).rename({0:'Region', 1:'Degree'}, axis=1)#.boxplot(by=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:06.733285Z",
     "start_time": "2020-08-10T19:20:06.729297Z"
    }
   },
   "outputs": [],
   "source": [
    "region_map = {'AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)':'ARMM',\n",
    "       'CORDILLERA ADMINISTRA TIVE REGION (CAR)':'CAR',\n",
    "       'REGION IV-B (MIMAROPA)':'IV-B', 'NATIONAL CAPITAL REGION (NCR)':'NCR',\n",
    "       'REGION X (NORTHERN MINDANAO)':'X', 'REGION XI (DAVAO REGION)':'XI',\n",
    "       'REGION XII (SOCCSKSA RGEN)':'XII', 'REGION XIII (CARAGA)':'XIII',\n",
    "       'REGION I (ILOCOS REGION)':'I', 'REGION II (CAGAYAN VALLEY)':'II',\n",
    "       'REGION III (CENTRAL LUZON)':'III', 'REGION IV-A (CALABAR ZON)':'IV-A',\n",
    "       'REGION V (BICOL REGION)':'V', 'REGION VI (WESTERN VISAYAS)':'VI',\n",
    "       'REGION VII (CENTRAL VISAYAS)':'VII', 'REGION VIII (EASTERN VISAYAS)':'VIII',\n",
    "       'REGION IX (ZAMBOANGA PENINSULA)':'IX'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:07.078180Z",
     "start_time": "2020-08-10T19:20:07.074190Z"
    }
   },
   "outputs": [],
   "source": [
    "df_deg['Region'] = df_deg['Region'].map(region_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:07.591288Z",
     "start_time": "2020-08-10T19:20:07.578324Z"
    }
   },
   "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>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Region</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ARMM</th>\n",
       "      <td>43</td>\n",
       "      <td>3.72</td>\n",
       "      <td>1.00</td>\n",
       "      <td>76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>27</td>\n",
       "      <td>7.54</td>\n",
       "      <td>6.00</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>58</td>\n",
       "      <td>9.66</td>\n",
       "      <td>7.00</td>\n",
       "      <td>148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>45</td>\n",
       "      <td>13.63</td>\n",
       "      <td>11.00</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>24</td>\n",
       "      <td>6.53</td>\n",
       "      <td>5.00</td>\n",
       "      <td>264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>60</td>\n",
       "      <td>9.97</td>\n",
       "      <td>8.00</td>\n",
       "      <td>183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>50</td>\n",
       "      <td>13.84</td>\n",
       "      <td>11.00</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>39</td>\n",
       "      <td>8.04</td>\n",
       "      <td>6.50</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>25</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.00</td>\n",
       "      <td>384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>40</td>\n",
       "      <td>9.53</td>\n",
       "      <td>8.00</td>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>47</td>\n",
       "      <td>12.17</td>\n",
       "      <td>10.00</td>\n",
       "      <td>142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>37</td>\n",
       "      <td>11.29</td>\n",
       "      <td>9.00</td>\n",
       "      <td>141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>22</td>\n",
       "      <td>4.84</td>\n",
       "      <td>4.00</td>\n",
       "      <td>161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>73</td>\n",
       "      <td>10.98</td>\n",
       "      <td>9.00</td>\n",
       "      <td>88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>41</td>\n",
       "      <td>15.94</td>\n",
       "      <td>16.00</td>\n",
       "      <td>63</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>49</td>\n",
       "      <td>16.78</td>\n",
       "      <td>18.00</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>35</td>\n",
       "      <td>4.95</td>\n",
       "      <td>3.50</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        max  mean  median  count\n",
       "Region                          \n",
       "ARMM     43  3.72    1.00     76\n",
       "CAR      27  7.54    6.00     98\n",
       "I        58  9.66    7.00    148\n",
       "II       45 13.63   11.00     86\n",
       "III      24  6.53    5.00    264\n",
       "IV-A     60  9.97    8.00    183\n",
       "IV-B     50 13.84   11.00     79\n",
       "IX       39  8.04    6.50     96\n",
       "NCR      25  0.29    0.00    384\n",
       "V        40  9.53    8.00    129\n",
       "VI       47 12.17   10.00    142\n",
       "VII      37 11.29    9.00    141\n",
       "VIII     22  4.84    4.00    161\n",
       "X        73 10.98    9.00     88\n",
       "XI       41 15.94   16.00     63\n",
       "XII      49 16.78   18.00     46\n",
       "XIII     35  4.95    3.50     58"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df_deg.groupby('Region')['Degree']\n",
    " .agg(['min', 'max', 'mean', 'median', 'count']).sort_values(by='Region')\n",
    " [['max', 'mean', 'median', 'count']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Degree Outliers (RHU)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:08.615976Z",
     "start_time": "2020-08-10T19:20:08.606987Z"
    }
   },
   "outputs": [],
   "source": [
    "upper_fence = df_deg.groupby('Region')['Degree'].quantile(0.75) + 1.5 * (df_deg.groupby('Region')['Degree'].quantile(0.75) - df_deg.groupby('Region')['Degree'].quantile(0.25))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:09.284885Z",
     "start_time": "2020-08-10T19:20:08.962746Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "\n",
    "# fig = plt.figure()\n",
    "sns.boxplot(data=df_deg, x='Region', y='Degree', palette='Blues',\n",
    "            order=upper_fence.sort_values(ascending=False).index)\n",
    "labels = [l.get_text() for l in ax.get_xticklabels()]\n",
    "ax.set_xticklabels(labels, ha='right');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:10.776343Z",
     "start_time": "2020-08-10T19:20:10.765373Z"
    }
   },
   "outputs": [],
   "source": [
    "df_deg['Outlierness'] = df_deg['Degree'] - df_deg['Region'].map(upper_fence)\n",
    "deg_outlier_rhu = df_deg[df_deg['Outlierness'] > 0].groupby('Region')['Outlierness'].agg(['min', 'max', 'mean', 'median', 'count'])\n",
    "deg_outlier_rhu.loc['CAR'] = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Average path length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:11.505653Z",
     "start_time": "2020-08-10T19:20:11.251198Z"
    }
   },
   "outputs": [],
   "source": [
    "sp = list(nx.all_pairs_dijkstra_path_length(G, weight='DIST'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:11.916285Z",
     "start_time": "2020-08-10T19:20:11.901302Z"
    }
   },
   "outputs": [],
   "source": [
    "sp = pd.DataFrame(sp).set_index(0)[1].to_dict()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### BHS -> RHU path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:25.089977Z",
     "start_time": "2020-08-10T19:20:12.312201Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs_rhu_ave_sp = {}\n",
    "for rhu_node in rhu_nodes:\n",
    "    ave_sp = 0\n",
    "    N = 0\n",
    "    for bhs_node in bhs_nodes:\n",
    "        if sp.get(bhs_node) and sp.get(bhs_node, {}).get(rhu_node):\n",
    "            ave_sp += sp[bhs_node][rhu_node]\n",
    "            N += 1\n",
    "    if N:\n",
    "        ave_sp /= N\n",
    "        bhs_rhu_ave_sp[rhu_node] = ave_sp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:25.484013Z",
     "start_time": "2020-08-10T19:20:25.478029Z"
    }
   },
   "outputs": [],
   "source": [
    "df_bhs_rhu = pd.DataFrame(bhs_rhu_ave_sp.items(), columns=['HF_CODE', 'AVE_SPL'])\n",
    "df_bhs_rhu['REGION'] = df_bhs_rhu['HF_CODE'].map(nodes['REGION']).map(region_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:25.870494Z",
     "start_time": "2020-08-10T19:20:25.864482Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs_rhu_spl = df_bhs_rhu.groupby('REGION')['AVE_SPL'].agg(['min', 'max', 'mean', 'median', 'count'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:26.259461Z",
     "start_time": "2020-08-10T19:20:26.251448Z"
    }
   },
   "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>min</th>\n",
       "      <th>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION</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>ARMM</th>\n",
       "      <td>0.65</td>\n",
       "      <td>30.09</td>\n",
       "      <td>5.56</td>\n",
       "      <td>4.06</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>1.52</td>\n",
       "      <td>21.32</td>\n",
       "      <td>6.02</td>\n",
       "      <td>5.03</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>0.00</td>\n",
       "      <td>31.66</td>\n",
       "      <td>4.45</td>\n",
       "      <td>3.59</td>\n",
       "      <td>140</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>2.39</td>\n",
       "      <td>31.66</td>\n",
       "      <td>5.75</td>\n",
       "      <td>5.40</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>0.09</td>\n",
       "      <td>24.48</td>\n",
       "      <td>3.19</td>\n",
       "      <td>2.70</td>\n",
       "      <td>237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>0.21</td>\n",
       "      <td>30.29</td>\n",
       "      <td>3.93</td>\n",
       "      <td>3.35</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>2.34</td>\n",
       "      <td>51.20</td>\n",
       "      <td>8.06</td>\n",
       "      <td>5.81</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>0.86</td>\n",
       "      <td>42.92</td>\n",
       "      <td>6.72</td>\n",
       "      <td>4.83</td>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>0.25</td>\n",
       "      <td>73.30</td>\n",
       "      <td>14.82</td>\n",
       "      <td>3.23</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>0.13</td>\n",
       "      <td>32.46</td>\n",
       "      <td>6.71</td>\n",
       "      <td>4.98</td>\n",
       "      <td>118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>0.54</td>\n",
       "      <td>38.20</td>\n",
       "      <td>5.56</td>\n",
       "      <td>4.35</td>\n",
       "      <td>136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>1.09</td>\n",
       "      <td>28.91</td>\n",
       "      <td>5.10</td>\n",
       "      <td>4.27</td>\n",
       "      <td>134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>0.04</td>\n",
       "      <td>86.02</td>\n",
       "      <td>6.28</td>\n",
       "      <td>4.21</td>\n",
       "      <td>145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>1.00</td>\n",
       "      <td>23.34</td>\n",
       "      <td>5.95</td>\n",
       "      <td>4.62</td>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>0.79</td>\n",
       "      <td>47.85</td>\n",
       "      <td>8.91</td>\n",
       "      <td>6.29</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>2.23</td>\n",
       "      <td>67.11</td>\n",
       "      <td>9.84</td>\n",
       "      <td>6.95</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>0.65</td>\n",
       "      <td>34.38</td>\n",
       "      <td>6.72</td>\n",
       "      <td>5.89</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        min   max  mean  median  count\n",
       "REGION                                \n",
       "ARMM   0.65 30.09  5.56    4.06     43\n",
       "CAR    1.52 21.32  6.02    5.03     75\n",
       "I      0.00 31.66  4.45    3.59    140\n",
       "II     2.39 31.66  5.75    5.40     82\n",
       "III    0.09 24.48  3.19    2.70    237\n",
       "IV-A   0.21 30.29  3.93    3.35    157\n",
       "IV-B   2.34 51.20  8.06    5.81     75\n",
       "IX     0.86 42.92  6.72    4.83     92\n",
       "NCR    0.25 73.30 14.82    3.23     19\n",
       "V      0.13 32.46  6.71    4.98    118\n",
       "VI     0.54 38.20  5.56    4.35    136\n",
       "VII    1.09 28.91  5.10    4.27    134\n",
       "VIII   0.04 86.02  6.28    4.21    145\n",
       "X      1.00 23.34  5.95    4.62     83\n",
       "XI     0.79 47.85  8.91    6.29     58\n",
       "XII    2.23 67.11  9.84    6.95     44\n",
       "XIII   0.65 34.38  6.72    5.89     43"
      ]
     },
     "execution_count": 198,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bhs_rhu_spl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Comparison with Degree Outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:26.797024Z",
     "start_time": "2020-08-10T19:20:26.646393Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "ax.plot(deg_outlier_rhu['mean'], bhs_rhu_spl['mean'], 'o', color='deepskyblue', \n",
    "        markersize='15')\n",
    "for i,j,k in zip(deg_outlier_rhu['mean'].values, bhs_rhu_spl['mean'].values, list(deg_outlier_rhu.index)):\n",
    "    ax.text(i,j,k, fontsize=14)\n",
    "ax.set_xlabel(\"Degree Outlierness (RHU)\", fontsize=14)\n",
    "ax.set_ylabel(\"Average shortest path length (BHS->RHU)\", fontsize=14);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### RHU -> HOSP path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:27.618794Z",
     "start_time": "2020-08-10T19:20:27.187945Z"
    }
   },
   "outputs": [],
   "source": [
    "rhu_hosp_ave_sp = {}\n",
    "for hosp_node in hosp_nodes:\n",
    "    ave_sp = 0\n",
    "    N = 0\n",
    "    for rhu_node in rhu_nodes:\n",
    "        if sp.get(rhu_node) and sp.get(rhu_node, {}).get(hosp_node):\n",
    "            ave_sp += sp[rhu_node][hosp_node]\n",
    "            N += 1\n",
    "    if N:\n",
    "        ave_sp /= N\n",
    "        rhu_hosp_ave_sp[hosp_node] = ave_sp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:28.035679Z",
     "start_time": "2020-08-10T19:20:28.028697Z"
    }
   },
   "outputs": [],
   "source": [
    "df_rhu_hosp = pd.DataFrame(rhu_hosp_ave_sp.items(), columns=['HF_CODE', 'AVE_SPL'])\n",
    "df_rhu_hosp['REGION'] = df_rhu_hosp['HF_CODE'].map(nodes['REGION']).map(region_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:28.429658Z",
     "start_time": "2020-08-10T19:20:28.419654Z"
    }
   },
   "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>min</th>\n",
       "      <th>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION</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>ARMM</th>\n",
       "      <td>13.96</td>\n",
       "      <td>52.74</td>\n",
       "      <td>30.15</td>\n",
       "      <td>32.53</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>6.52</td>\n",
       "      <td>30.51</td>\n",
       "      <td>16.47</td>\n",
       "      <td>16.00</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>5.73</td>\n",
       "      <td>34.93</td>\n",
       "      <td>16.85</td>\n",
       "      <td>14.64</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>0.22</td>\n",
       "      <td>43.62</td>\n",
       "      <td>18.39</td>\n",
       "      <td>14.65</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>3.53</td>\n",
       "      <td>33.86</td>\n",
       "      <td>12.97</td>\n",
       "      <td>10.79</td>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>3.13</td>\n",
       "      <td>72.94</td>\n",
       "      <td>16.03</td>\n",
       "      <td>15.33</td>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>5.94</td>\n",
       "      <td>54.63</td>\n",
       "      <td>27.00</td>\n",
       "      <td>25.51</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>0.42</td>\n",
       "      <td>56.19</td>\n",
       "      <td>23.80</td>\n",
       "      <td>22.07</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>1.20</td>\n",
       "      <td>33.93</td>\n",
       "      <td>7.21</td>\n",
       "      <td>3.62</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>10.95</td>\n",
       "      <td>52.01</td>\n",
       "      <td>22.74</td>\n",
       "      <td>19.48</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>7.50</td>\n",
       "      <td>40.18</td>\n",
       "      <td>18.40</td>\n",
       "      <td>16.46</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>5.43</td>\n",
       "      <td>45.77</td>\n",
       "      <td>20.58</td>\n",
       "      <td>18.71</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>7.55</td>\n",
       "      <td>66.16</td>\n",
       "      <td>26.83</td>\n",
       "      <td>23.98</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>1.53</td>\n",
       "      <td>53.28</td>\n",
       "      <td>20.28</td>\n",
       "      <td>17.49</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>7.14</td>\n",
       "      <td>75.00</td>\n",
       "      <td>29.56</td>\n",
       "      <td>25.20</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>14.24</td>\n",
       "      <td>81.92</td>\n",
       "      <td>34.14</td>\n",
       "      <td>19.55</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>7.66</td>\n",
       "      <td>40.07</td>\n",
       "      <td>19.27</td>\n",
       "      <td>18.85</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         min   max  mean  median  count\n",
       "REGION                                 \n",
       "ARMM   13.96 52.74 30.15   32.53     18\n",
       "CAR     6.52 30.51 16.47   16.00     35\n",
       "I       5.73 34.93 16.85   14.64     36\n",
       "II      0.22 43.62 18.39   14.65     32\n",
       "III     3.53 33.86 12.97   10.79     57\n",
       "IV-A    3.13 72.94 16.03   15.33     57\n",
       "IV-B    5.94 54.63 27.00   25.51     27\n",
       "IX      0.42 56.19 23.80   22.07     21\n",
       "NCR     1.20 33.93  7.21    3.62     45\n",
       "V      10.95 52.01 22.74   19.48     39\n",
       "VI      7.50 40.18 18.40   16.46     47\n",
       "VII     5.43 45.77 20.58   18.71     45\n",
       "VIII    7.55 66.16 26.83   23.98     39\n",
       "X       1.53 53.28 20.28   17.49     36\n",
       "XI      7.14 75.00 29.56   25.20     20\n",
       "XII    14.24 81.92 34.14   19.55     17\n",
       "XIII    7.66 40.07 19.27   18.85     25"
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rhu_hosp_spl = df_rhu_hosp.groupby('REGION')['AVE_SPL'].agg(['min', 'max', 'mean', 'median', 'count'])\n",
    "rhu_hosp_spl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Comparison with Degree Outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:28.962273Z",
     "start_time": "2020-08-10T19:20:28.814596Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "ax.plot(deg_outlier['mean'], rhu_hosp_spl['mean'], 'o', color='deepskyblue', \n",
    "        markersize='15')\n",
    "for i,j,k in zip(deg_outlier['mean'].values, rhu_hosp_spl['mean'].values, \n",
    "                 list(deg_outlier.index)):\n",
    "    ax.text(i,j,k, fontsize=14)\n",
    "ax.set_xlabel(\"Degree Outlierness (Hosp)\", fontsize=14)\n",
    "ax.set_ylabel(\"Average shortest path length (RHU->Hosp)\", fontsize=14);\n",
    "\n",
    "#I changed the ylabel to RHU->Hosp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### BHS -> HOSP path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:33.833376Z",
     "start_time": "2020-08-10T19:20:29.367157Z"
    }
   },
   "outputs": [],
   "source": [
    "bhs_hosp_ave_sp = {}\n",
    "for hosp_node in hosp_nodes:\n",
    "    ave_sp = 0\n",
    "    N = 0\n",
    "    for bhs_node in bhs_nodes:\n",
    "        if sp.get(bhs_node) and sp.get(bhs_node, {}).get(hosp_node):\n",
    "            ave_sp += sp[bhs_node][hosp_node]\n",
    "            N += 1\n",
    "    if N:\n",
    "        ave_sp /= N\n",
    "        bhs_hosp_ave_sp[hosp_node] = ave_sp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:34.225428Z",
     "start_time": "2020-08-10T19:20:34.219478Z"
    }
   },
   "outputs": [],
   "source": [
    "df_bhs_hosp = pd.DataFrame(bhs_hosp_ave_sp.items(), columns=['HF_CODE', 'AVE_SPL'])\n",
    "df_bhs_hosp['REGION'] = df_bhs_hosp['HF_CODE'].map(nodes['REGION']).map(region_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:34.624362Z",
     "start_time": "2020-08-10T19:20:34.612445Z"
    }
   },
   "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>min</th>\n",
       "      <th>max</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION</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>ARMM</th>\n",
       "      <td>4.51</td>\n",
       "      <td>65.32</td>\n",
       "      <td>35.23</td>\n",
       "      <td>38.25</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CAR</th>\n",
       "      <td>10.50</td>\n",
       "      <td>40.53</td>\n",
       "      <td>22.46</td>\n",
       "      <td>21.81</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>I</th>\n",
       "      <td>3.78</td>\n",
       "      <td>36.87</td>\n",
       "      <td>18.46</td>\n",
       "      <td>16.85</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>II</th>\n",
       "      <td>4.40</td>\n",
       "      <td>49.01</td>\n",
       "      <td>21.31</td>\n",
       "      <td>21.14</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>III</th>\n",
       "      <td>1.27</td>\n",
       "      <td>37.93</td>\n",
       "      <td>15.56</td>\n",
       "      <td>12.76</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-A</th>\n",
       "      <td>1.00</td>\n",
       "      <td>78.27</td>\n",
       "      <td>18.08</td>\n",
       "      <td>15.30</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IV-B</th>\n",
       "      <td>3.78</td>\n",
       "      <td>56.02</td>\n",
       "      <td>29.10</td>\n",
       "      <td>28.42</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IX</th>\n",
       "      <td>0.87</td>\n",
       "      <td>61.03</td>\n",
       "      <td>25.67</td>\n",
       "      <td>24.71</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NCR</th>\n",
       "      <td>1.21</td>\n",
       "      <td>68.62</td>\n",
       "      <td>19.19</td>\n",
       "      <td>6.63</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V</th>\n",
       "      <td>0.67</td>\n",
       "      <td>51.68</td>\n",
       "      <td>25.13</td>\n",
       "      <td>22.92</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VI</th>\n",
       "      <td>9.23</td>\n",
       "      <td>49.10</td>\n",
       "      <td>21.48</td>\n",
       "      <td>19.70</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VII</th>\n",
       "      <td>6.58</td>\n",
       "      <td>62.52</td>\n",
       "      <td>23.95</td>\n",
       "      <td>23.19</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VIII</th>\n",
       "      <td>12.82</td>\n",
       "      <td>88.35</td>\n",
       "      <td>33.74</td>\n",
       "      <td>29.56</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>X</th>\n",
       "      <td>7.83</td>\n",
       "      <td>52.84</td>\n",
       "      <td>23.69</td>\n",
       "      <td>22.38</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XI</th>\n",
       "      <td>1.86</td>\n",
       "      <td>74.25</td>\n",
       "      <td>29.39</td>\n",
       "      <td>30.16</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XII</th>\n",
       "      <td>2.80</td>\n",
       "      <td>71.17</td>\n",
       "      <td>30.50</td>\n",
       "      <td>23.27</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XIII</th>\n",
       "      <td>6.83</td>\n",
       "      <td>57.35</td>\n",
       "      <td>24.26</td>\n",
       "      <td>24.00</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         min   max  mean  median  count\n",
       "REGION                                 \n",
       "ARMM    4.51 65.32 35.23   38.25     17\n",
       "CAR    10.50 40.53 22.46   21.81     35\n",
       "I       3.78 36.87 18.46   16.85     39\n",
       "II      4.40 49.01 21.31   21.14     35\n",
       "III     1.27 37.93 15.56   12.76     62\n",
       "IV-A    1.00 78.27 18.08   15.30     60\n",
       "IV-B    3.78 56.02 29.10   28.42     30\n",
       "IX      0.87 61.03 25.67   24.71     23\n",
       "NCR     1.21 68.62 19.19    6.63     41\n",
       "V       0.67 51.68 25.13   22.92     41\n",
       "VI      9.23 49.10 21.48   19.70     48\n",
       "VII     6.58 62.52 23.95   23.19     48\n",
       "VIII   12.82 88.35 33.74   29.56     39\n",
       "X       7.83 52.84 23.69   22.38     36\n",
       "XI      1.86 74.25 29.39   30.16     25\n",
       "XII     2.80 71.17 30.50   23.27     20\n",
       "XIII    6.83 57.35 24.26   24.00     26"
      ]
     },
     "execution_count": 206,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bhs_hosp_spl = df_bhs_hosp.groupby('REGION')['AVE_SPL'].agg(['min', 'max', 'mean', 'median', 'count'])\n",
    "bhs_hosp_spl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Comparison with Degree Outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:35.170946Z",
     "start_time": "2020-08-10T19:20:35.013322Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "ax.plot(deg_outlier['mean'], bhs_hosp_spl['mean'], 'o', color='deepskyblue', \n",
    "        markersize='15')\n",
    "for i,j,k in zip(deg_outlier['mean'].values, bhs_hosp_spl['mean'].values, list(deg_outlier.index)):\n",
    "    ax.text(i,j,k, fontsize=14)\n",
    "ax.set_xlabel(\"Degree Outlierness (Hosp)\", fontsize=14)\n",
    "ax.set_ylabel(\"Average shortest path length (BHS->Hosp)\", fontsize=14);\n",
    "\n",
    "\n",
    "## I changed the ylabel to BHS -> Hosp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simulations\n",
    "After characterizing the network, further analysis were performed by simulating what would happen to the network as higher percentages of the population enters the system. That is, the facilities that would overload are identified based on the bed capacity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:35.574323Z",
     "start_time": "2020-08-10T19:20:35.558331Z"
    }
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</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>HF_TYPE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BARANGAY HEALTH STATION</th>\n",
       "      <td>17,435.00</td>\n",
       "      <td>3,800.18</td>\n",
       "      <td>6,013.85</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1,302.00</td>\n",
       "      <td>2,444.00</td>\n",
       "      <td>4,408.50</td>\n",
       "      <td>123,708.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOSPITAL</th>\n",
       "      <td>0.00</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>INFIRMARY</th>\n",
       "      <td>0.00</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RURAL HEALTH UNIT</th>\n",
       "      <td>2,172.00</td>\n",
       "      <td>38,663.14</td>\n",
       "      <td>44,947.38</td>\n",
       "      <td>0.00</td>\n",
       "      <td>16,981.75</td>\n",
       "      <td>30,224.00</td>\n",
       "      <td>46,929.25</td>\n",
       "      <td>903,309.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            count      mean       std  min       25%  \\\n",
       "HF_TYPE                                                                \n",
       "BARANGAY HEALTH STATION 17,435.00  3,800.18  6,013.85 0.00  1,302.00   \n",
       "HOSPITAL                     0.00       nan       nan  nan       nan   \n",
       "INFIRMARY                    0.00       nan       nan  nan       nan   \n",
       "RURAL HEALTH UNIT        2,172.00 38,663.14 44,947.38 0.00 16,981.75   \n",
       "\n",
       "                              50%       75%        max  \n",
       "HF_TYPE                                                 \n",
       "BARANGAY HEALTH STATION  2,444.00  4,408.50 123,708.00  \n",
       "HOSPITAL                      nan       nan        nan  \n",
       "INFIRMARY                     nan       nan        nan  \n",
       "RURAL HEALTH UNIT       30,224.00 46,929.25 903,309.00  "
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nodes.groupby('HF_TYPE')['CATCHMENT'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:35.977328Z",
     "start_time": "2020-08-10T19:20:35.965339Z"
    }
   },
   "outputs": [],
   "source": [
    "intake = nodes.loc[(nodes['HF_TYPE']=='RURAL HEALTH UNIT') | (nodes['HF_TYPE']=='BARANGAY HEALTH STATION'), ['REGION', 'CATCHMENT', 'HF_TYPE']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:36.390253Z",
     "start_time": "2020-08-10T19:20:36.377336Z"
    }
   },
   "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>median</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)</th>\n",
       "      <td>9,252.73</td>\n",
       "      <td>2,682.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>222,885.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CORDILLERA ADMINISTRA TIVE REGION (CAR)</th>\n",
       "      <td>3,688.14</td>\n",
       "      <td>1,187.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>134,461.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NATIONAL CAPITAL REGION (NCR)</th>\n",
       "      <td>30,745.29</td>\n",
       "      <td>25,084.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>290,977.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION I (ILOCOS REGION)</th>\n",
       "      <td>6,133.94</td>\n",
       "      <td>2,556.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>204,135.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION II (CAGAYAN VALLEY)</th>\n",
       "      <td>4,528.67</td>\n",
       "      <td>1,826.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>158,245.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION III (CENTRAL LUZON)</th>\n",
       "      <td>9,866.27</td>\n",
       "      <td>3,700.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>260,691.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-A (CALABAR ZON)</th>\n",
       "      <td>9,040.42</td>\n",
       "      <td>3,122.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>463,727.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-B (MIMAROPA)</th>\n",
       "      <td>5,363.06</td>\n",
       "      <td>2,104.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>272,190.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IX (ZAMBOANGA PENINSULA)</th>\n",
       "      <td>9,427.20</td>\n",
       "      <td>3,773.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>903,309.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION V (BICOL REGION)</th>\n",
       "      <td>7,685.29</td>\n",
       "      <td>3,480.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>210,047.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VI (WESTERN VISAYAS)</th>\n",
       "      <td>7,620.31</td>\n",
       "      <td>3,293.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>597,740.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VII (CENTRAL VISAYAS)</th>\n",
       "      <td>5,827.44</td>\n",
       "      <td>2,131.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>414,265.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VIII (EASTERN VISAYAS)</th>\n",
       "      <td>9,092.89</td>\n",
       "      <td>4,577.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>387,813.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION X (NORTHERN MINDANAO)</th>\n",
       "      <td>6,047.59</td>\n",
       "      <td>2,369.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>197,890.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XI (DAVAO REGION)</th>\n",
       "      <td>5,155.10</td>\n",
       "      <td>1,818.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>175,935.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XII (SOCCSKSA RGEN)</th>\n",
       "      <td>8,722.27</td>\n",
       "      <td>2,168.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>453,585.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XIII (CARAGA)</th>\n",
       "      <td>7,212.89</td>\n",
       "      <td>2,530.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>137,527.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 mean    median  min  \\\n",
       "REGION                                                                 \n",
       "AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)  9,252.73  2,682.00 0.00   \n",
       "CORDILLERA ADMINISTRA TIVE REGION (CAR)      3,688.14  1,187.00 0.00   \n",
       "NATIONAL CAPITAL REGION (NCR)               30,745.29 25,084.00 0.00   \n",
       "REGION I (ILOCOS REGION)                     6,133.94  2,556.00 0.00   \n",
       "REGION II (CAGAYAN VALLEY)                   4,528.67  1,826.00 0.00   \n",
       "REGION III (CENTRAL LUZON)                   9,866.27  3,700.50 0.00   \n",
       "REGION IV-A (CALABAR ZON)                    9,040.42  3,122.50 0.00   \n",
       "REGION IV-B (MIMAROPA)                       5,363.06  2,104.50 0.00   \n",
       "REGION IX (ZAMBOANGA PENINSULA)              9,427.20  3,773.50 0.00   \n",
       "REGION V (BICOL REGION)                      7,685.29  3,480.50 0.00   \n",
       "REGION VI (WESTERN VISAYAS)                  7,620.31  3,293.00 0.00   \n",
       "REGION VII (CENTRAL VISAYAS)                 5,827.44  2,131.00 0.00   \n",
       "REGION VIII (EASTERN VISAYAS)                9,092.89  4,577.00 0.00   \n",
       "REGION X (NORTHERN MINDANAO)                 6,047.59  2,369.00 0.00   \n",
       "REGION XI (DAVAO REGION)                     5,155.10  1,818.50 0.00   \n",
       "REGION XII (SOCCSKSA RGEN)                   8,722.27  2,168.00 0.00   \n",
       "REGION XIII (CARAGA)                         7,212.89  2,530.00 0.00   \n",
       "\n",
       "                                                   max  \n",
       "REGION                                                  \n",
       "AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM) 222,885.00  \n",
       "CORDILLERA ADMINISTRA TIVE REGION (CAR)     134,461.00  \n",
       "NATIONAL CAPITAL REGION (NCR)               290,977.00  \n",
       "REGION I (ILOCOS REGION)                    204,135.00  \n",
       "REGION II (CAGAYAN VALLEY)                  158,245.00  \n",
       "REGION III (CENTRAL LUZON)                  260,691.00  \n",
       "REGION IV-A (CALABAR ZON)                   463,727.00  \n",
       "REGION IV-B (MIMAROPA)                      272,190.00  \n",
       "REGION IX (ZAMBOANGA PENINSULA)             903,309.00  \n",
       "REGION V (BICOL REGION)                     210,047.00  \n",
       "REGION VI (WESTERN VISAYAS)                 597,740.00  \n",
       "REGION VII (CENTRAL VISAYAS)                414,265.00  \n",
       "REGION VIII (EASTERN VISAYAS)               387,813.00  \n",
       "REGION X (NORTHERN MINDANAO)                197,890.00  \n",
       "REGION XI (DAVAO REGION)                    175,935.00  \n",
       "REGION XII (SOCCSKSA RGEN)                  453,585.00  \n",
       "REGION XIII (CARAGA)                        137,527.00  "
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "intake.groupby('REGION')['CATCHMENT'].agg(['mean','median', 'min', 'max'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:36.796167Z",
     "start_time": "2020-08-10T19:20:36.788190Z"
    }
   },
   "outputs": [],
   "source": [
    "num_iter=1000\n",
    "#check changes in the system if 0.001% to 1% of population enters the HCPN;\n",
    "pop_prop = 0.00001\n",
    "hosp_zero_cap = nodes[((nodes['HF_TYPE']=='HOSPITAL') | (nodes['HF_TYPE']=='INFIRMARY')) & (nodes['BED_CAP']==0)].SHORT_HFCODE.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:37.206071Z",
     "start_time": "2020-08-10T19:20:37.202081Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hosp_zero_cap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:20:49.529374Z",
     "start_time": "2020-08-10T19:20:37.607258Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Percentage of isolated RHUs')"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, [ax1, ax2] = plt.subplots(1, 2, figsize=(9, 6))\n",
    "\n",
    "intake_t = intake['CATCHMENT'] * pop_prop\n",
    "df_isolates = pd.DataFrame()\n",
    "df_rem_hosps = pd.DataFrame()\n",
    "# ax2 = ax1.twinx()\n",
    "for region in ['NATIONAL CAPITAL REGION (NCR)', 'REGION VIII (EASTERN VISAYAS)']:\n",
    "    rhu_reg = nodes[(nodes['REGION']==region) & (nodes['HF_TYPE']=='RURAL HEALTH UNIT')]['SHORT_HFCODE']\n",
    "    hosp_reg = nodes[(nodes['REGION']==region) & ((nodes['HF_TYPE']=='HOSPITAL') | (nodes['HF_TYPE']=='INFIRMARY'))]['SHORT_HFCODE']\n",
    "    G_reg = G.subgraph(rhu_reg.tolist() + hosp_reg.tolist()).copy()\n",
    "    hosp_reg_connected = [i for i in hosp_reg if ((i not in nx.isolates(G_reg)) and (i not in hosp_zero_cap))]\n",
    "    remaining_hosps = [len(hosp_reg_connected)]\n",
    "    G_reg.remove_nodes_from(list(nx.isolates(G_reg)))\n",
    "    G_reg.remove_nodes_from(hosp_zero_cap)\n",
    "    num_isolates = [0]\n",
    "    remaining_hosp = [i for i, j in G_reg.nodes(data=True) if ((j['HF_TYPE']=='HOSPITAL') or (j['HF_TYPE']=='INFIRMARY'))]\n",
    "    # t=0\n",
    "    # while remaining_hosp:\n",
    "    for t in range(num_iter):\n",
    "        removed_hosps = []\n",
    "        for rhu in rhu_reg:\n",
    "            try:\n",
    "                N = len(list(G_reg.neighbors(rhu)))\n",
    "            except nx.NetworkXError:\n",
    "                continue\n",
    "            nearby_hosps = list(G_reg.neighbors(rhu))\n",
    "#             dist_sum = G_reg.out_degree(rhu, weight='DIST')\n",
    "            dist_sum = sum([G_reg[rhu][i]['DIST'] for i in nearby_hosps if G_reg[rhu][i]['DIST']==G_reg[rhu][i]['DIST']])\n",
    "            for hosp in nearby_hosps:\n",
    "                if len(nearby_hosps)==1:\n",
    "                    prob_hosp = 1\n",
    "                else:\n",
    "                    prob_hosp = (1 - G_reg[rhu][hosp]['DIST'] / dist_sum) if G_reg[rhu][hosp]['DIST']==G_reg[rhu][hosp]['DIST'] else 0\n",
    "                new_patient = intake_t[rhu] * prob_hosp if intake_t[rhu]==intake_t[rhu] else 0\n",
    "                add_one = 1 if random.random()<new_patient%1 else 0\n",
    "                if (G_reg.nodes()[hosp].get('patients', 0) + int(new_patient) + add_one) > G_reg.nodes()[hosp]['BED_CAP']:\n",
    "                    G_reg.remove_node(hosp)\n",
    "                    removed_hosps.append(hosp)\n",
    "                else:\n",
    "                    nx.set_node_attributes(G_reg, {hosp: G_reg.nodes()[hosp].get('patients', 0) + int(new_patient) + add_one}, 'patients')\n",
    "        remaining_hosp = [i for i, j in G_reg.nodes(data=True) if ((j['HF_TYPE']=='HOSPITAL') or (j['HF_TYPE']=='INFIRMARY'))]\n",
    "        remaining_hosps.append(len(remaining_hosp))\n",
    "        num_isolates.append(len(list(nx.isolates(G_reg))))\n",
    "    #     print(t, removed_hosps)\n",
    "    df_rem_hosps.loc[region, \"0.05%\"] = remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / remaining_hosps[0]\n",
    "    df_rem_hosps.loc[region, \"0.5%\"] = remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / remaining_hosps[0]\n",
    "    df_isolates.loc[region, \"0.05%\"] = num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / num_isolates[-1]\n",
    "    df_isolates.loc[region, \"0.5%\"] = num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / num_isolates[-1]\n",
    "    \n",
    "#     print(\"At 0.05%, remaining hosps are\", remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / remaining_hosps[0], region)\n",
    "#     print(\"At 0.5%, remaining hosps are\", remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / remaining_hosps[0], region)\n",
    "#     print(\"At 0.05%, the isolated RHUs are\", num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / num_isolates[-1], region)\n",
    "#     print(\"At 0.5%, the isolated RHUs are\", num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / num_isolates[-1], region)\n",
    "    ax1.plot(pop_prop * np.arange(num_iter+1), [i/remaining_hosps[0] for i in remaining_hosps], label=region)\n",
    "    ax2.plot(pop_prop * np.arange(num_iter+1), [i/num_isolates[-1] for i in num_isolates])\n",
    "fig.legend()\n",
    "ax1.set_xlabel('Percent of population in HCPN')\n",
    "ax2.set_xlabel('Percent of population in HCPN')\n",
    "\n",
    "ax1.set_xticklabels(['-0.2%', '0.0%', '0.2%', '0.4%', '0.6%', '0.8%', '1.0%'])\n",
    "ax2.set_xticklabels(['-0.2%', '0.0%', '0.2%', '0.4%', '0.6%', '0.8%', '1.0%'])\n",
    "ax1.set_ylabel('Percentage of remaining hospitals')\n",
    "ax2.set_ylabel('Percentage of isolated RHUs')\n",
    "    #     t+=1\n",
    "    #     break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:21:26.223286Z",
     "start_time": "2020-08-10T19:20:49.930271Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Percentage of isolated RHUs')"
      ]
     },
     "execution_count": 214,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, [ax1, ax2] = plt.subplots(1, 2, figsize=(9, 6))\n",
    "\n",
    "intake_t = intake['CATCHMENT'] * pop_prop\n",
    "df_isolates = pd.DataFrame()\n",
    "df_rem_hosps = pd.DataFrame()\n",
    "# ax2 = ax1.twinx()\n",
    "for region in nodes['REGION'].unique():\n",
    "    rhu_reg = nodes[(nodes['REGION']==region) & (nodes['HF_TYPE']=='RURAL HEALTH UNIT')]['SHORT_HFCODE']\n",
    "    hosp_reg = nodes[(nodes['REGION']==region) & ((nodes['HF_TYPE']=='HOSPITAL') | (nodes['HF_TYPE']=='INFIRMARY'))]['SHORT_HFCODE']\n",
    "    G_reg = G.subgraph(rhu_reg.tolist() + hosp_reg.tolist()).copy()\n",
    "    hosp_reg_connected = [i for i in hosp_reg if ((i not in nx.isolates(G_reg)) and (i not in hosp_zero_cap))]\n",
    "    remaining_hosps = [len(hosp_reg_connected)]\n",
    "    G_reg.remove_nodes_from(list(nx.isolates(G_reg)))\n",
    "    G_reg.remove_nodes_from(hosp_zero_cap)\n",
    "    num_isolates = [0]\n",
    "    remaining_hosp = [i for i, j in G_reg.nodes(data=True) if ((j['HF_TYPE']=='HOSPITAL') or (j['HF_TYPE']=='INFIRMARY'))]\n",
    "    # t=0\n",
    "    # while remaining_hosp:\n",
    "    for t in range(num_iter):\n",
    "        removed_hosps = []\n",
    "        for rhu in rhu_reg:\n",
    "            try:\n",
    "                N = len(list(G_reg.neighbors(rhu)))\n",
    "            except nx.NetworkXError:\n",
    "                continue\n",
    "            nearby_hosps = list(G_reg.neighbors(rhu))\n",
    "#             dist_sum = G_reg.out_degree(rhu, weight='DIST')\n",
    "            dist_sum = sum([G_reg[rhu][i]['DIST'] for i in nearby_hosps if G_reg[rhu][i]['DIST']==G_reg[rhu][i]['DIST']])\n",
    "            for hosp in nearby_hosps:\n",
    "                if len(nearby_hosps)==1:\n",
    "                    prob_hosp = 1\n",
    "                else:\n",
    "                    prob_hosp = (1 - G_reg[rhu][hosp]['DIST'] / dist_sum) if G_reg[rhu][hosp]['DIST']==G_reg[rhu][hosp]['DIST'] else 0\n",
    "                new_patient = intake_t[rhu] * prob_hosp if intake_t[rhu]==intake_t[rhu] else 0\n",
    "                add_one = 1 if random.random()<new_patient%1 else 0\n",
    "                if (G_reg.nodes()[hosp].get('patients', 0) + int(new_patient) + add_one) > G_reg.nodes()[hosp]['BED_CAP']:\n",
    "                    G_reg.remove_node(hosp)\n",
    "                    removed_hosps.append(hosp)\n",
    "                else:\n",
    "                    nx.set_node_attributes(G_reg, {hosp: G_reg.nodes()[hosp].get('patients', 0) + int(new_patient) + add_one}, 'patients')\n",
    "        remaining_hosp = [i for i, j in G_reg.nodes(data=True) if ((j['HF_TYPE']=='HOSPITAL') or (j['HF_TYPE']=='INFIRMARY'))]\n",
    "        remaining_hosps.append(len(remaining_hosp))\n",
    "        num_isolates.append(len(list(nx.isolates(G_reg))))\n",
    "    #     print(t, removed_hosps)\n",
    "    df_rem_hosps.loc[region, \"0.05%\"] = remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / remaining_hosps[0]\n",
    "    df_rem_hosps.loc[region, \"0.5%\"] = remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / remaining_hosps[0]\n",
    "    df_isolates.loc[region, \"0.05%\"] = num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / num_isolates[-1]\n",
    "    df_isolates.loc[region, \"0.5%\"] = num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / num_isolates[-1]\n",
    "    \n",
    "#     print(\"At 0.05%, remaining hosps are\", remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / remaining_hosps[0], region)\n",
    "#     print(\"At 0.5%, remaining hosps are\", remaining_hosps[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / remaining_hosps[0], region)\n",
    "#     print(\"At 0.05%, the isolated RHUs are\", num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.0005)[0][0]] / num_isolates[-1], region)\n",
    "#     print(\"At 0.5%, the isolated RHUs are\", num_isolates[np.where(pop_prop * np.arange(num_iter+1)==0.005)[0][0]] / num_isolates[-1], region)\n",
    "    ax1.plot(pop_prop * np.arange(num_iter+1), [i/remaining_hosps[0] for i in remaining_hosps], label=region)\n",
    "    ax2.plot(pop_prop * np.arange(num_iter+1), [i/num_isolates[-1] for i in num_isolates])\n",
    "fig.legend()\n",
    "ax1.set_xlabel('Percent of population in HCPN')\n",
    "ax2.set_xlabel('Percent of population in HCPN')\n",
    "\n",
    "ax1.set_xticklabels(['-0.2%', '0.0%', '0.2%', '0.4%', '0.6%', '0.8%', '1.0%'])\n",
    "ax2.set_xticklabels(['-0.2%', '0.0%', '0.2%', '0.4%', '0.6%', '0.8%', '1.0%'])\n",
    "ax1.set_ylabel('Percentage of remaining hospitals')\n",
    "ax2.set_ylabel('Percentage of isolated RHUs')\n",
    "    #     t+=1\n",
    "    #     break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:21:26.636182Z",
     "start_time": "2020-08-10T19:21:26.629169Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Remaining hospitals\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>0.05%</th>\n",
       "      <th>0.5%</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)</th>\n",
       "      <td>27.78</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CORDILLERA ADMINISTRA TIVE REGION (CAR)</th>\n",
       "      <td>41.67</td>\n",
       "      <td>2.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-B (MIMAROPA)</th>\n",
       "      <td>24.24</td>\n",
       "      <td>3.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NATIONAL CAPITAL REGION (NCR)</th>\n",
       "      <td>26.67</td>\n",
       "      <td>8.89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION X (NORTHERN MINDANAO)</th>\n",
       "      <td>22.22</td>\n",
       "      <td>5.56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XI (DAVAO REGION)</th>\n",
       "      <td>38.10</td>\n",
       "      <td>23.81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XII (SOCCSKSA RGEN)</th>\n",
       "      <td>38.89</td>\n",
       "      <td>22.22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XIII (CARAGA)</th>\n",
       "      <td>32.00</td>\n",
       "      <td>12.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION I (ILOCOS REGION)</th>\n",
       "      <td>13.89</td>\n",
       "      <td>5.56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION II (CAGAYAN VALLEY)</th>\n",
       "      <td>9.09</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION III (CENTRAL LUZON)</th>\n",
       "      <td>5.26</td>\n",
       "      <td>1.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-A (CALABAR ZON)</th>\n",
       "      <td>8.77</td>\n",
       "      <td>1.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION V (BICOL REGION)</th>\n",
       "      <td>5.13</td>\n",
       "      <td>2.56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VI (WESTERN VISAYAS)</th>\n",
       "      <td>8.00</td>\n",
       "      <td>2.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VII (CENTRAL VISAYAS)</th>\n",
       "      <td>11.11</td>\n",
       "      <td>4.44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VIII (EASTERN VISAYAS)</th>\n",
       "      <td>17.95</td>\n",
       "      <td>5.13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IX (ZAMBOANGA PENINSULA)</th>\n",
       "      <td>19.05</td>\n",
       "      <td>14.29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             0.05%  0.5%\n",
       "AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)  27.78  0.00\n",
       "CORDILLERA ADMINISTRA TIVE REGION (CAR)      41.67  2.78\n",
       "REGION IV-B (MIMAROPA)                       24.24  3.03\n",
       "NATIONAL CAPITAL REGION (NCR)                26.67  8.89\n",
       "REGION X (NORTHERN MINDANAO)                 22.22  5.56\n",
       "REGION XI (DAVAO REGION)                     38.10 23.81\n",
       "REGION XII (SOCCSKSA RGEN)                   38.89 22.22\n",
       "REGION XIII (CARAGA)                         32.00 12.00\n",
       "REGION I (ILOCOS REGION)                     13.89  5.56\n",
       "REGION II (CAGAYAN VALLEY)                    9.09  0.00\n",
       "REGION III (CENTRAL LUZON)                    5.26  1.75\n",
       "REGION IV-A (CALABAR ZON)                     8.77  1.75\n",
       "REGION V (BICOL REGION)                       5.13  2.56\n",
       "REGION VI (WESTERN VISAYAS)                   8.00  2.00\n",
       "REGION VII (CENTRAL VISAYAS)                 11.11  4.44\n",
       "REGION VIII (EASTERN VISAYAS)                17.95  5.13\n",
       "REGION IX (ZAMBOANGA PENINSULA)              19.05 14.29"
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Remaining hospitals\")\n",
    "(df_rem_hosps*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-10T19:21:27.035117Z",
     "start_time": "2020-08-10T19:21:27.028103Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RHU isolates\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>0.05%</th>\n",
       "      <th>0.5%</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)</th>\n",
       "      <td>56.60</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CORDILLERA ADMINISTRA TIVE REGION (CAR)</th>\n",
       "      <td>17.81</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-B (MIMAROPA)</th>\n",
       "      <td>76.09</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NATIONAL CAPITAL REGION (NCR)</th>\n",
       "      <td>48.15</td>\n",
       "      <td>92.59</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION X (NORTHERN MINDANAO)</th>\n",
       "      <td>52.63</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XI (DAVAO REGION)</th>\n",
       "      <td>5.26</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XII (SOCCSKSA RGEN)</th>\n",
       "      <td>40.00</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION XIII (CARAGA)</th>\n",
       "      <td>68.57</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION I (ILOCOS REGION)</th>\n",
       "      <td>96.12</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION II (CAGAYAN VALLEY)</th>\n",
       "      <td>79.49</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION III (CENTRAL LUZON)</th>\n",
       "      <td>94.42</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IV-A (CALABAR ZON)</th>\n",
       "      <td>91.11</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION V (BICOL REGION)</th>\n",
       "      <td>82.35</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VI (WESTERN VISAYAS)</th>\n",
       "      <td>79.79</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VII (CENTRAL VISAYAS)</th>\n",
       "      <td>97.47</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION VIII (EASTERN VISAYAS)</th>\n",
       "      <td>92.59</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION IX (ZAMBOANGA PENINSULA)</th>\n",
       "      <td>100.00</td>\n",
       "      <td>100.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             0.05%   0.5%\n",
       "AUTONOMOUS REGION IN MUSLIM MINDANAO (ARMM)  56.60 100.00\n",
       "CORDILLERA ADMINISTRA TIVE REGION (CAR)      17.81 100.00\n",
       "REGION IV-B (MIMAROPA)                       76.09 100.00\n",
       "NATIONAL CAPITAL REGION (NCR)                48.15  92.59\n",
       "REGION X (NORTHERN MINDANAO)                 52.63 100.00\n",
       "REGION XI (DAVAO REGION)                      5.26 100.00\n",
       "REGION XII (SOCCSKSA RGEN)                   40.00 100.00\n",
       "REGION XIII (CARAGA)                         68.57 100.00\n",
       "REGION I (ILOCOS REGION)                     96.12 100.00\n",
       "REGION II (CAGAYAN VALLEY)                   79.49 100.00\n",
       "REGION III (CENTRAL LUZON)                   94.42 100.00\n",
       "REGION IV-A (CALABAR ZON)                    91.11 100.00\n",
       "REGION V (BICOL REGION)                      82.35 100.00\n",
       "REGION VI (WESTERN VISAYAS)                  79.79 100.00\n",
       "REGION VII (CENTRAL VISAYAS)                 97.47 100.00\n",
       "REGION VIII (EASTERN VISAYAS)                92.59 100.00\n",
       "REGION IX (ZAMBOANGA PENINSULA)             100.00 100.00"
      ]
     },
     "execution_count": 216,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"RHU isolates\")\n",
    "(df_isolates*100)"
   ]
  }
 ],
 "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": "3.7.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}