age_fm_demo_pm.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fairness Metric Age Effect Differences\n",
    "This demonstration downloads census data by race/ethnicity and age.  Then, using \"older\" and \"younger\" groups, shows metric differences between racial/ethnic groups due to age, even when the underlying process does not depend on race/ethnicity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import censusdata\n",
    "import re\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy.interpolate import interp1d\n",
    "from scipy.special import expit\n",
    "import math\n",
    "import os\n",
    "import random\n",
    "from scipy.stats import gaussian_kde\n",
    "from textwrap import wrap"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Set local path for saving charts, etc.\n",
    "Input your path here!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "local_save_path = '/YOUR/PATH/HERE'\n",
    "os.path.isdir(local_save_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Census Data - Age by Race/Ethnicity\n",
    "I will analyze age effects on fairness metrics using the US population distributions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Census metadata\n",
    "Here, I get variable info for tables that I want to examine.  I am interested in reace/ethnicity by age information from the 2019 American Community Survey.  Information about tables I want to examine can be found here:\n",
    " - https://censusreporter.org/topics/age-sex/\n",
    " - https://censusreporter.org/topics/race-hispanic/\n",
    "\n",
    "In this section, I get metadata about variables I want to examine, which are needed for count pulls."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# List the tables I want to pul. \n",
    "table_list = ['B01001B', 'B01001C', 'B01001D', 'B01001E', 'B01001F', 'B01001G', 'B01001H', 'B01001I']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get metadata about fields in all the tables I want to use\n",
    "table_metdata_raw = pd.concat(map(lambda x: pd.DataFrame(censusdata.censustable('acs5', 2019, x)), \n",
    "                                        table_list), axis=1) \\\n",
    "    .transpose() \\\n",
    "    .reset_index() \\\n",
    "    .rename(columns={'index':'variable'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parse concept - get race/ethnicity category\n",
    "table_info_1_df= table_metdata_raw.copy()\n",
    "table_info_1_df['race_eth'] = table_metdata_raw.concept.str.split('[()]',2, expand=True).iloc[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parse label - get sex, age strings\n",
    "table_info_1_df['sex'] = table_info_1_df.label.str.split('[!!]', expand=True).iloc[:,4]\n",
    "table_info_1_df['age_grp'] = table_info_1_df.label.str.split('[!!]', expand=True).iloc[:,6]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop columns with missing sex, age groups\n",
    "table_info_1_df.dropna(how='any', subset=['sex', 'age_grp'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>variable</th>\n",
       "      <th>label</th>\n",
       "      <th>concept</th>\n",
       "      <th>predicateType</th>\n",
       "      <th>race_eth</th>\n",
       "      <th>sex</th>\n",
       "      <th>age_grp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B01001B_003E</td>\n",
       "      <td>Estimate!!Total:!!Male:!!Under 5 years</td>\n",
       "      <td>SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)</td>\n",
       "      <td>int</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>Under 5 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B01001B_004E</td>\n",
       "      <td>Estimate!!Total:!!Male:!!5 to 9 years</td>\n",
       "      <td>SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)</td>\n",
       "      <td>int</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>5 to 9 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B01001B_005E</td>\n",
       "      <td>Estimate!!Total:!!Male:!!10 to 14 years</td>\n",
       "      <td>SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)</td>\n",
       "      <td>int</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>10 to 14 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>B01001B_006E</td>\n",
       "      <td>Estimate!!Total:!!Male:!!15 to 17 years</td>\n",
       "      <td>SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)</td>\n",
       "      <td>int</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>15 to 17 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>B01001B_007E</td>\n",
       "      <td>Estimate!!Total:!!Male:!!18 and 19 years</td>\n",
       "      <td>SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)</td>\n",
       "      <td>int</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>18 and 19 years</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       variable                                     label  \\\n",
       "2  B01001B_003E    Estimate!!Total:!!Male:!!Under 5 years   \n",
       "3  B01001B_004E     Estimate!!Total:!!Male:!!5 to 9 years   \n",
       "4  B01001B_005E   Estimate!!Total:!!Male:!!10 to 14 years   \n",
       "5  B01001B_006E   Estimate!!Total:!!Male:!!15 to 17 years   \n",
       "6  B01001B_007E  Estimate!!Total:!!Male:!!18 and 19 years   \n",
       "\n",
       "                                        concept predicateType  \\\n",
       "2  SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)           int   \n",
       "3  SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)           int   \n",
       "4  SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)           int   \n",
       "5  SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)           int   \n",
       "6  SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE)           int   \n",
       "\n",
       "                          race_eth    sex          age_grp  \n",
       "2  BLACK OR AFRICAN AMERICAN ALONE  Male:    Under 5 years  \n",
       "3  BLACK OR AFRICAN AMERICAN ALONE  Male:     5 to 9 years  \n",
       "4  BLACK OR AFRICAN AMERICAN ALONE  Male:   10 to 14 years  \n",
       "5  BLACK OR AFRICAN AMERICAN ALONE  Male:   15 to 17 years  \n",
       "6  BLACK OR AFRICAN AMERICAN ALONE  Male:  18 and 19 years  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table_info_1_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Census age group string parsing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Handle the lowest age group\n",
    "age_tbl_1 = table_info_1_df.age_grp.str.extract(r'Under (?P<age_max>[0-9]*) years').dropna().astype('int')\n",
    "age_tbl_1['age_min'] = 0\n",
    "age_tbl_1['age_max'] = age_tbl_1['age_max'] - 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Handle the highest age group\n",
    "age_tbl_2 = table_info_1_df.age_grp.str.extract(r'(?P<age_min>[0-9]*) years and over').dropna().astype('int')\n",
    "age_tbl_2['age_max'] = 99 # arbitray end point for plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age_min</th>\n",
       "      <th>age_max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>15</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>18</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>20</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>242</th>\n",
       "      <td>35</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>45</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>244</th>\n",
       "      <td>55</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245</th>\n",
       "      <td>65</td>\n",
       "      <td>74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>246</th>\n",
       "      <td>75</td>\n",
       "      <td>84</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>192 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     age_min  age_max\n",
       "3          5        9\n",
       "4         10       14\n",
       "5         15       17\n",
       "6         18       19\n",
       "7         20       24\n",
       "..       ...      ...\n",
       "242       35       44\n",
       "243       45       54\n",
       "244       55       64\n",
       "245       65       74\n",
       "246       75       84\n",
       "\n",
       "[192 rows x 2 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Handle age ranges \n",
    "age_tbl_3 = table_info_1_df.age_grp.str.extract(r'(?P<age_min>[0-9]*) (?:to|and) (?P<age_max>[0-9]*) years').dropna().astype('int')\n",
    "age_tbl_3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Handle 2 year ranges\n",
    "age_tbl_4 = table_info_1_df.age_grp.str.extract(r'(?P<age_min>[0-9]*) and (?P<age_max>[0-9]*) years').dropna().astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Combine the info\n",
    "age_tbl_comb = age_tbl_3.combine_first(age_tbl_2).combine_first(age_tbl_1).astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "table_info_2_df = table_info_1_df.join(age_tbl_comb, how='left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Census field conversions\n",
    "Convert some fields to categorical, drop others"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "table_info_df = table_info_2_df.copy()[['variable', 'race_eth', 'sex', 'age_grp', 'age_min', 'age_max']]\n",
    "table_info_df['race_eth'] = pd.Categorical(table_info_2_df.race_eth, categories=table_info_2_df.race_eth.unique())\n",
    "table_info_df['age_grp'] = pd.Categorical(table_info_2_df.age_grp, categories=table_info_2_df.age_grp.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>variable</th>\n",
       "      <th>race_eth</th>\n",
       "      <th>sex</th>\n",
       "      <th>age_grp</th>\n",
       "      <th>age_min</th>\n",
       "      <th>age_max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B01001B_003E</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>Under 5 years</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B01001B_004E</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>5 to 9 years</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B01001B_005E</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>10 to 14 years</td>\n",
       "      <td>10</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>B01001B_006E</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>15 to 17 years</td>\n",
       "      <td>15</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>B01001B_007E</td>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Male:</td>\n",
       "      <td>18 and 19 years</td>\n",
       "      <td>18</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       variable                         race_eth    sex          age_grp  \\\n",
       "2  B01001B_003E  BLACK OR AFRICAN AMERICAN ALONE  Male:    Under 5 years   \n",
       "3  B01001B_004E  BLACK OR AFRICAN AMERICAN ALONE  Male:     5 to 9 years   \n",
       "4  B01001B_005E  BLACK OR AFRICAN AMERICAN ALONE  Male:   10 to 14 years   \n",
       "5  B01001B_006E  BLACK OR AFRICAN AMERICAN ALONE  Male:   15 to 17 years   \n",
       "6  B01001B_007E  BLACK OR AFRICAN AMERICAN ALONE  Male:  18 and 19 years   \n",
       "\n",
       "   age_min  age_max  \n",
       "2        0        4  \n",
       "3        5        9  \n",
       "4       10       14  \n",
       "5       15       17  \n",
       "6       18       19  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table_info_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Census population counts by age with race/ethnicity\n",
    "Here, I pull the variables identified above and aggregate to combine sex categories."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>B01001B_003E</th>\n",
       "      <th>B01001B_004E</th>\n",
       "      <th>B01001B_005E</th>\n",
       "      <th>B01001B_006E</th>\n",
       "      <th>B01001B_007E</th>\n",
       "      <th>B01001B_008E</th>\n",
       "      <th>B01001B_009E</th>\n",
       "      <th>B01001B_010E</th>\n",
       "      <th>B01001B_011E</th>\n",
       "      <th>B01001B_012E</th>\n",
       "      <th>...</th>\n",
       "      <th>B01001I_022E</th>\n",
       "      <th>B01001I_023E</th>\n",
       "      <th>B01001I_024E</th>\n",
       "      <th>B01001I_025E</th>\n",
       "      <th>B01001I_026E</th>\n",
       "      <th>B01001I_027E</th>\n",
       "      <th>B01001I_028E</th>\n",
       "      <th>B01001I_029E</th>\n",
       "      <th>B01001I_030E</th>\n",
       "      <th>B01001I_031E</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>United States: Summary level: 010, us:1</th>\n",
       "      <td>1390883</td>\n",
       "      <td>1432012</td>\n",
       "      <td>1485845</td>\n",
       "      <td>911088</td>\n",
       "      <td>644594</td>\n",
       "      <td>1662549</td>\n",
       "      <td>1678054</td>\n",
       "      <td>1406596</td>\n",
       "      <td>2520532</td>\n",
       "      <td>2476388</td>\n",
       "      <td>...</td>\n",
       "      <td>954620</td>\n",
       "      <td>2321468</td>\n",
       "      <td>2262448</td>\n",
       "      <td>2140552</td>\n",
       "      <td>4093134</td>\n",
       "      <td>3373036</td>\n",
       "      <td>2396670</td>\n",
       "      <td>1378777</td>\n",
       "      <td>695641</td>\n",
       "      <td>286388</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 224 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         B01001B_003E  B01001B_004E  \\\n",
       "United States: Summary level: 010, us:1       1390883       1432012   \n",
       "\n",
       "                                         B01001B_005E  B01001B_006E  \\\n",
       "United States: Summary level: 010, us:1       1485845        911088   \n",
       "\n",
       "                                         B01001B_007E  B01001B_008E  \\\n",
       "United States: Summary level: 010, us:1        644594       1662549   \n",
       "\n",
       "                                         B01001B_009E  B01001B_010E  \\\n",
       "United States: Summary level: 010, us:1       1678054       1406596   \n",
       "\n",
       "                                         B01001B_011E  B01001B_012E  ...  \\\n",
       "United States: Summary level: 010, us:1       2520532       2476388  ...   \n",
       "\n",
       "                                         B01001I_022E  B01001I_023E  \\\n",
       "United States: Summary level: 010, us:1        954620       2321468   \n",
       "\n",
       "                                         B01001I_024E  B01001I_025E  \\\n",
       "United States: Summary level: 010, us:1       2262448       2140552   \n",
       "\n",
       "                                         B01001I_026E  B01001I_027E  \\\n",
       "United States: Summary level: 010, us:1       4093134       3373036   \n",
       "\n",
       "                                         B01001I_028E  B01001I_029E  \\\n",
       "United States: Summary level: 010, us:1       2396670       1378777   \n",
       "\n",
       "                                         B01001I_030E  B01001I_031E  \n",
       "United States: Summary level: 010, us:1        695641        286388  \n",
       "\n",
       "[1 rows x 224 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Pull the entire US counts for all variables listed in table_info_df\n",
    "census_counts_df = censusdata.download('acs5', 2019,\n",
    "                             censusdata.censusgeo([('us','*')]),\n",
    "                             table_info_df.variable.to_list())\n",
    "census_counts_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>variable</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>B01001B_003E</td>\n",
       "      <td>1390883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B01001B_004E</td>\n",
       "      <td>1432012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B01001B_005E</td>\n",
       "      <td>1485845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B01001B_006E</td>\n",
       "      <td>911088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B01001B_007E</td>\n",
       "      <td>644594</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       variable    count\n",
       "0  B01001B_003E  1390883\n",
       "1  B01001B_004E  1432012\n",
       "2  B01001B_005E  1485845\n",
       "3  B01001B_006E   911088\n",
       "4  B01001B_007E   644594"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Convert to long format\n",
    "census_long_df = census_counts_df.transpose().astype('int').reset_index()\n",
    "census_long_df.columns= ['variable', 'count']\n",
    "census_long_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Join in metadata\n",
    "census_long_df = census_long_df.merge(table_info_df[['variable', 'race_eth', 'sex', 'age_grp', 'age_min', 'age_max']], \n",
    "                                                    how='inner', on='variable')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_grp</th>\n",
       "      <th>count</th>\n",
       "      <th>age_min</th>\n",
       "      <th>age_max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>Under 5 years</td>\n",
       "      <td>2740124</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>5 to 9 years</td>\n",
       "      <td>2828410</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>10 to 14 years</td>\n",
       "      <td>2929886</td>\n",
       "      <td>10</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>15 to 17 years</td>\n",
       "      <td>1786717</td>\n",
       "      <td>15</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BLACK OR AFRICAN AMERICAN ALONE</td>\n",
       "      <td>18 and 19 years</td>\n",
       "      <td>1264392</td>\n",
       "      <td>18</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          race_eth          age_grp    count  age_min  age_max\n",
       "0  BLACK OR AFRICAN AMERICAN ALONE    Under 5 years  2740124        0        4\n",
       "1  BLACK OR AFRICAN AMERICAN ALONE     5 to 9 years  2828410        5        9\n",
       "2  BLACK OR AFRICAN AMERICAN ALONE   10 to 14 years  2929886       10       14\n",
       "3  BLACK OR AFRICAN AMERICAN ALONE   15 to 17 years  1786717       15       17\n",
       "4  BLACK OR AFRICAN AMERICAN ALONE  18 and 19 years  1264392       18       19"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Aggregate to combine sex categories\n",
    "census_long_comb_df = census_long_df.groupby(['race_eth', 'age_grp']) \\\n",
    "    .agg({'count':np.sum, 'age_min': 'first', 'age_max':'first'}) \\\n",
    "    .reset_index()\n",
    "census_long_comb_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "85 years and over    8\n",
       "75 to 84 years       8\n",
       "65 to 74 years       8\n",
       "55 to 64 years       8\n",
       "45 to 54 years       8\n",
       "35 to 44 years       8\n",
       "30 to 34 years       8\n",
       "25 to 29 years       8\n",
       "20 to 24 years       8\n",
       "18 and 19 years      8\n",
       "15 to 17 years       8\n",
       "10 to 14 years       8\n",
       "5 to 9 years         8\n",
       "Under 5 years        8\n",
       "Name: age_grp, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "census_long_comb_df['age_grp'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Census age interpolation (approximate age distribution)\n",
    "Here, I go from discrete age bins to a dataframe that has 1 row per race/ethnicity and age in years.  This is done via interpolation and so results in an approximate smoothed shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply the counts to the bin midpoint - scale to account for differing bin sizes\n",
    "census_re_1 =  census_long_comb_df.copy()\n",
    "census_re_1['age_mid'] = (census_re_1['age_max'] + census_re_1['age_max']) / 2\n",
    "census_re_1['age_width'] = census_re_1['age_max'] - census_re_1['age_min'] + 1\n",
    "census_re_1['count_scale'] = census_re_1['count'] / census_re_1['age_width']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Write a function to interpolate to age level\n",
    "def get_age_interp(data):\n",
    "    xnew = np.linspace(np.min(data.age_min), np.max(data.age_max), \n",
    "            num = math.ceil(1 + np.max(data.age_max) - np.min(data.age_min)), endpoint=True)\n",
    "    f_cubic = interp1d(data['age_mid'], data['count_scale'], kind='quadratic',\n",
    "                  fill_value = 'extrapolate')\n",
    "    return pd.DataFrame({'age_yr': xnew, 'count_interp':f_cubic(xnew)})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to each race/ethnicity group\n",
    "census_interpolated_dist = census_re_1.groupby('race_eth').apply(get_age_interp).reset_index()\n",
    "census_interpolated_dist.drop(columns='level_1', errors='ignore', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get relative weights within groups\n",
    "census_interpolated_dist['count_weight'] = census_interpolated_dist.groupby('race_eth')['count_interp'].transform(lambda x: x / np.sum(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Plot approximate age distributions by race/ethnicity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>race_eth</th>\n",
       "      <th>BLACK OR AFRICAN AMERICAN ALONE</th>\n",
       "      <th>AMERICAN INDIAN AND ALASKA NATIVE ALONE</th>\n",
       "      <th>ASIAN ALONE</th>\n",
       "      <th>NATIVE HAWAIIAN AND OTHER PACIFIC ISLANDER ALONE</th>\n",
       "      <th>SOME OTHER RACE ALONE</th>\n",
       "      <th>TWO OR MORE RACES</th>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age_yr</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>0.0</th>\n",
       "      <td>0.012344</td>\n",
       "      <td>0.012374</td>\n",
       "      <td>0.008721</td>\n",
       "      <td>0.013460</td>\n",
       "      <td>0.014966</td>\n",
       "      <td>0.028183</td>\n",
       "      <td>0.009579</td>\n",
       "      <td>0.016932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>0.012378</td>\n",
       "      <td>0.012553</td>\n",
       "      <td>0.008969</td>\n",
       "      <td>0.013334</td>\n",
       "      <td>0.014827</td>\n",
       "      <td>0.027751</td>\n",
       "      <td>0.009525</td>\n",
       "      <td>0.016706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.0</th>\n",
       "      <td>0.012421</td>\n",
       "      <td>0.012734</td>\n",
       "      <td>0.009199</td>\n",
       "      <td>0.013225</td>\n",
       "      <td>0.014728</td>\n",
       "      <td>0.027317</td>\n",
       "      <td>0.009488</td>\n",
       "      <td>0.016520</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3.0</th>\n",
       "      <td>0.012471</td>\n",
       "      <td>0.012916</td>\n",
       "      <td>0.009413</td>\n",
       "      <td>0.013131</td>\n",
       "      <td>0.014669</td>\n",
       "      <td>0.026882</td>\n",
       "      <td>0.009467</td>\n",
       "      <td>0.016375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4.0</th>\n",
       "      <td>0.012528</td>\n",
       "      <td>0.013099</td>\n",
       "      <td>0.009609</td>\n",
       "      <td>0.013054</td>\n",
       "      <td>0.014649</td>\n",
       "      <td>0.026446</td>\n",
       "      <td>0.009463</td>\n",
       "      <td>0.016271</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race_eth  BLACK OR AFRICAN AMERICAN ALONE  \\\n",
       "age_yr                                      \n",
       "0.0                              0.012344   \n",
       "1.0                              0.012378   \n",
       "2.0                              0.012421   \n",
       "3.0                              0.012471   \n",
       "4.0                              0.012528   \n",
       "\n",
       "race_eth  AMERICAN INDIAN AND ALASKA NATIVE ALONE  ASIAN ALONE  \\\n",
       "age_yr                                                           \n",
       "0.0                                      0.012374     0.008721   \n",
       "1.0                                      0.012553     0.008969   \n",
       "2.0                                      0.012734     0.009199   \n",
       "3.0                                      0.012916     0.009413   \n",
       "4.0                                      0.013099     0.009609   \n",
       "\n",
       "race_eth  NATIVE HAWAIIAN AND OTHER PACIFIC ISLANDER ALONE  \\\n",
       "age_yr                                                       \n",
       "0.0                                               0.013460   \n",
       "1.0                                               0.013334   \n",
       "2.0                                               0.013225   \n",
       "3.0                                               0.013131   \n",
       "4.0                                               0.013054   \n",
       "\n",
       "race_eth  SOME OTHER RACE ALONE  TWO OR MORE RACES  \\\n",
       "age_yr                                               \n",
       "0.0                    0.014966           0.028183   \n",
       "1.0                    0.014827           0.027751   \n",
       "2.0                    0.014728           0.027317   \n",
       "3.0                    0.014669           0.026882   \n",
       "4.0                    0.014649           0.026446   \n",
       "\n",
       "race_eth  WHITE ALONE, NOT HISPANIC OR LATINO  HISPANIC OR LATINO  \n",
       "age_yr                                                             \n",
       "0.0                                  0.009579            0.016932  \n",
       "1.0                                  0.009525            0.016706  \n",
       "2.0                                  0.009488            0.016520  \n",
       "3.0                                  0.009467            0.016375  \n",
       "4.0                                  0.009463            0.016271  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Pivot to get groups in columns\n",
    "pivot_dist = pd.pivot_table(census_interpolated_dist,\n",
    "               index='age_yr', columns='race_eth', values='count_weight')\n",
    "pivot_dist.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot grid info\n",
    "num_plots = len(pivot_dist.columns)\n",
    "num_rows = math.ceil(num_plots/2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create the plots\n",
    "fig, axes = plt.subplots(num_rows, 2, sharex=True, sharey=False, figsize=(10,10))\n",
    "for i, ser_nm in enumerate(pivot_dist.columns):\n",
    "    # Plot data, fit\n",
    "    ax1 = axes[math.floor(i/2), i % 2]\n",
    "    ax1.plot(pivot_dist.index, pivot_dist[ser_nm], color=sns.color_palette().as_hex()[i])\n",
    "    ax1.title.set_text(ser_nm)\n",
    "    ax1.get_yaxis().set_visible(False)\n",
    "# Create a large frame to label x axis\n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "ax.set_xlabel('age', labelpad=10, fontsize=14) \n",
    "plt.tight_layout()\n",
    "plt.savefig(os.path.join(local_save_path, 'PLOT_age_dist_race_ethnicity.png'))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>race_eth</th>\n",
       "      <th>BLACK OR AFRICAN AMERICAN ALONE</th>\n",
       "      <th>AMERICAN INDIAN AND ALASKA NATIVE ALONE</th>\n",
       "      <th>ASIAN ALONE</th>\n",
       "      <th>NATIVE HAWAIIAN AND OTHER PACIFIC ISLANDER ALONE</th>\n",
       "      <th>SOME OTHER RACE ALONE</th>\n",
       "      <th>TWO OR MORE RACES</th>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age_yr</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>0.0</th>\n",
       "      <td>0.012344</td>\n",
       "      <td>0.012374</td>\n",
       "      <td>0.008721</td>\n",
       "      <td>0.013460</td>\n",
       "      <td>0.014966</td>\n",
       "      <td>0.028183</td>\n",
       "      <td>0.009579</td>\n",
       "      <td>0.016932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>0.012378</td>\n",
       "      <td>0.012553</td>\n",
       "      <td>0.008969</td>\n",
       "      <td>0.013334</td>\n",
       "      <td>0.014827</td>\n",
       "      <td>0.027751</td>\n",
       "      <td>0.009525</td>\n",
       "      <td>0.016706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.0</th>\n",
       "      <td>0.012421</td>\n",
       "      <td>0.012734</td>\n",
       "      <td>0.009199</td>\n",
       "      <td>0.013225</td>\n",
       "      <td>0.014728</td>\n",
       "      <td>0.027317</td>\n",
       "      <td>0.009488</td>\n",
       "      <td>0.016520</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3.0</th>\n",
       "      <td>0.012471</td>\n",
       "      <td>0.012916</td>\n",
       "      <td>0.009413</td>\n",
       "      <td>0.013131</td>\n",
       "      <td>0.014669</td>\n",
       "      <td>0.026882</td>\n",
       "      <td>0.009467</td>\n",
       "      <td>0.016375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4.0</th>\n",
       "      <td>0.012528</td>\n",
       "      <td>0.013099</td>\n",
       "      <td>0.009609</td>\n",
       "      <td>0.013054</td>\n",
       "      <td>0.014649</td>\n",
       "      <td>0.026446</td>\n",
       "      <td>0.009463</td>\n",
       "      <td>0.016271</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race_eth  BLACK OR AFRICAN AMERICAN ALONE  \\\n",
       "age_yr                                      \n",
       "0.0                              0.012344   \n",
       "1.0                              0.012378   \n",
       "2.0                              0.012421   \n",
       "3.0                              0.012471   \n",
       "4.0                              0.012528   \n",
       "\n",
       "race_eth  AMERICAN INDIAN AND ALASKA NATIVE ALONE  ASIAN ALONE  \\\n",
       "age_yr                                                           \n",
       "0.0                                      0.012374     0.008721   \n",
       "1.0                                      0.012553     0.008969   \n",
       "2.0                                      0.012734     0.009199   \n",
       "3.0                                      0.012916     0.009413   \n",
       "4.0                                      0.013099     0.009609   \n",
       "\n",
       "race_eth  NATIVE HAWAIIAN AND OTHER PACIFIC ISLANDER ALONE  \\\n",
       "age_yr                                                       \n",
       "0.0                                               0.013460   \n",
       "1.0                                               0.013334   \n",
       "2.0                                               0.013225   \n",
       "3.0                                               0.013131   \n",
       "4.0                                               0.013054   \n",
       "\n",
       "race_eth  SOME OTHER RACE ALONE  TWO OR MORE RACES  \\\n",
       "age_yr                                               \n",
       "0.0                    0.014966           0.028183   \n",
       "1.0                    0.014827           0.027751   \n",
       "2.0                    0.014728           0.027317   \n",
       "3.0                    0.014669           0.026882   \n",
       "4.0                    0.014649           0.026446   \n",
       "\n",
       "race_eth  WHITE ALONE, NOT HISPANIC OR LATINO  HISPANIC OR LATINO  \n",
       "age_yr                                                             \n",
       "0.0                                  0.009579            0.016932  \n",
       "1.0                                  0.009525            0.016706  \n",
       "2.0                                  0.009488            0.016520  \n",
       "3.0                                  0.009467            0.016375  \n",
       "4.0                                  0.009463            0.016271  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pivot_dist.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot only the 2 groups of interest\n",
    "pivot_dist_lim = pivot_dist.drop(columns=['BLACK OR AFRICAN AMERICAN ALONE',\n",
    "                                         'AMERICAN INDIAN AND ALASKA NATIVE ALONE',\n",
    "                                         'ASIAN ALONE',\n",
    "                                         'NATIVE HAWAIIAN AND OTHER PACIFIC ISLANDER ALONE',\n",
    "                                         'SOME OTHER RACE ALONE', 'TWO OR MORE RACES'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 2, sharex=True, sharey=False, figsize=(10, 3))\n",
    "for i, ser_nm in enumerate(pivot_dist_lim.columns):\n",
    "    # Plot data, fit\n",
    "    ax1 = axes[i]\n",
    "    ax1.plot(pivot_dist_lim.index, pivot_dist_lim[ser_nm])\n",
    "    ax1.title.set_text(ser_nm)\n",
    "    ax1.get_yaxis().set_visible(False)\n",
    "# Create a large frame to label x axis\n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "ax.set_xlabel('age', labelpad=10, fontsize=14) \n",
    "plt.tight_layout()\n",
    "plt.savefig(os.path.join(local_save_path, 'PLOT_age_dist_race_ethnicity_lim.png'))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Demonstration Data Preparation\n",
    "Simulate process that is age-dependent but independent of race/ethnicity.  Use the age 18+ population."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Draw 5,000 each random ages from 2 distributions (White Alone, not Hispanic or Latino and Hispanic or Latino)\n",
    "- Calculate age-independent random feature \"x\", as well as noise/random feature \"eta\"\n",
    "- Get target probability for each case - incorporate age, \"x\" and \"eta\"\n",
    "- Get target status based on probability"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Draw 5,000 each random ages from 2 distributions (White Alone, not Hispanic or Latino and Hispanic or Latino)\n",
    "I simulate a sample of individuals in two race/ethnicity groups.  I choose the oldest and one of the youngest race/ethnicity groups to show maximum effect of the age difference.  Restrict to the 18+ population.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get age 18+ dists for the selected race/ethnicity groups\n",
    "census_selected = census_interpolated_dist[census_interpolated_dist['race_eth'] \\\n",
    "                                           .isin(['WHITE ALONE, NOT HISPANIC OR LATINO',\n",
    "                                                 'HISPANIC OR LATINO']) &\n",
    "                                          (census_interpolated_dist['age_yr'] >=18)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to draw a random sample from the above distributions\n",
    "def sample_age_weighted(data, n=1000,\n",
    "                       weight_var='count_weight'):\n",
    "    \"\"\" Draw sample from an age distribution given in data.\n",
    "    Require the data have cases and weights with nonzero sum\"\"\"\n",
    "    if (len(data) == 0) or np.sum(data[weight_var] == 0):\n",
    "        return pd.DataFrame()\n",
    "    else:\n",
    "        return data.sample(n=n, weights=data[weight_var], replace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>52.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>69.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth  age_yr\n",
       "0  WHITE ALONE, NOT HISPANIC OR LATINO    55.0\n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO    55.0\n",
       "2  WHITE ALONE, NOT HISPANIC OR LATINO    52.0\n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO    20.0\n",
       "4  WHITE ALONE, NOT HISPANIC OR LATINO    69.0"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_data = census_selected[['race_eth', 'age_yr', 'count_weight']].groupby('race_eth') \\\n",
    "    .apply(lambda x: sample_age_weighted(x, 5000)) \\\n",
    "    .drop(columns='count_weight') \\\n",
    "    .reset_index(drop=True)\n",
    "pop_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x7f39bdcf26a0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x7f39bda7f208>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Quick check of the age distributions\n",
    "pop_data['age_yr'].hist(by=pop_data['race_eth'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "age_yr    47.7857\n",
      "dtype: float64\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>age_yr</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "      <td>43.6982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <td>51.8732</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      age_yr\n",
       "race_eth                                    \n",
       "HISPANIC OR LATINO                   43.6982\n",
       "WHITE ALONE, NOT HISPANIC OR LATINO  51.8732"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Show overall mean age, group age\n",
    "print(pop_data.agg({'age_yr':'mean'}))\n",
    "pop_data.groupby('race_eth').agg({'age_yr':'mean'})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate age-independent random feature \"x\", as well as noise/random feature \"eta\"\n",
    "I add an age-independent feature (x) as well as a noise feature to the data.  These are completely random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "pop_data['x'] = np.random.normal(size = pop_data.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "pop_data['eta'] = np.random.normal(size = pop_data.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "      <th>x</th>\n",
       "      <th>eta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>1.624345</td>\n",
       "      <td>-0.122474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>-0.611756</td>\n",
       "      <td>0.228170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>52.0</td>\n",
       "      <td>-0.528172</td>\n",
       "      <td>-0.352305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>20.0</td>\n",
       "      <td>-1.072969</td>\n",
       "      <td>-0.830553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>69.0</td>\n",
       "      <td>0.865408</td>\n",
       "      <td>-0.261090</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth  age_yr         x       eta\n",
       "0  WHITE ALONE, NOT HISPANIC OR LATINO    55.0  1.624345 -0.122474\n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO    55.0 -0.611756  0.228170\n",
       "2  WHITE ALONE, NOT HISPANIC OR LATINO    52.0 -0.528172 -0.352305\n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO    20.0 -1.072969 -0.830553\n",
       "4  WHITE ALONE, NOT HISPANIC OR LATINO    69.0  0.865408 -0.261090"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get target probability for each case - incorporate age, \"x\" and \"eta\"\n",
    "\"x\" and \"age\" have roughly similar contributions, eta is much smaller.  I don't standardize the age feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to get linear equation y = age + x + eta\n",
    "def lineq(age, x, eta, coef_int = -6,\n",
    "          coef_age = 1/10, coef_x = 1, coef_eta = 0.1):\n",
    "    return (coef_int + coef_age * age + \n",
    "            coef_x * x  + \n",
    "            coef_eta * eta)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add the linear equation to the age\n",
    "pop_data['sim_lin'] = pop_data.apply(lambda x: lineq(x.age_yr, x.x, x.eta), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Simulate the probability based on the linear equation\n",
    "pop_data['sim_prob'] = pop_data['sim_lin'].apply(expit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Check probability averages, distributions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "sim_prob    0.329845\n",
       "dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_data.agg({'sim_prob':'mean'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sim_prob</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "      <td>0.266758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <td>0.392932</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     sim_prob\n",
       "race_eth                                     \n",
       "HISPANIC OR LATINO                   0.266758\n",
       "WHITE ALONE, NOT HISPANIC OR LATINO  0.392932"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_data.groupby('race_eth').agg({'sim_prob':'mean'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x7f39bd9be2b0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x7f39bdf9a240>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pop_data['sim_prob'].hist(by=pop_data['race_eth'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get target status based on probability\n",
    "Note that the rate of target status for white alone, not hispanic or latino is nearly double hispanic or latino, based on the strong age feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "pop_data['targ_status'] = np.random.binomial(1, pop_data['sim_lin'].apply(expit), \n",
    "                                          size=pop_data.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>targ_status</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "      <td>0.2764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <td>0.3880</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     targ_status\n",
       "race_eth                                        \n",
       "HISPANIC OR LATINO                        0.2764\n",
       "WHITE ALONE, NOT HISPANIC OR LATINO       0.3880"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_data.groupby('race_eth').agg({'targ_status': 'mean'})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression Modeling\n",
    "Use the data to get two logistic regression models:\n",
    "- y = x + age\n",
    "- y = x\n",
    "\n",
    "For the first model, we should see fit coefficients similar to our inputs. \n",
    "\n",
    "For both, *show model probability distributions and 0.5 threshold*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_x_age = pop_data[['x', 'age_yr']]\n",
    "X_x = pop_data[['x']]\n",
    "y = pop_data['targ_status']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fit 2 models\n",
    "clf_x_age = LogisticRegression(random_state=0,  fit_intercept=True).fit(X_x_age, y)\n",
    "clf_x = LogisticRegression(random_state=0,  fit_intercept=True).fit(X_x, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   model  intercept         x       age\n",
      "0  x_age  -6.018028  1.064978  0.100525\n",
      "1      x  -0.769743  0.651085       NaN\n"
     ]
    }
   ],
   "source": [
    "# Coefficients to data frame\n",
    "coef_df = pd.DataFrame.from_dict({'model': ['x_age', 'x'],\n",
    "                        'intercept': [clf_x_age.intercept_[0], clf_x.intercept_[0]],\n",
    "                        'x': [clf_x_age.coef_[0,0], clf_x.coef_[0,0]],\n",
    "                        'age': [clf_x_age.coef_[0,1], None]})\n",
    "print(coef_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the coefficients\n",
    "coef_df.to_csv(os.path.join(local_save_path, 'DATA_regression_fit_coef.csv'), index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate predictions - probability and normal scales.  Save to a new data frame\n",
    "prediction_df = pd.DataFrame.from_dict({'prob_x_age': clf_x_age.predict_proba(X_x_age)[:,1],\n",
    "                                        'prob_x': clf_x.predict_proba(X_x)[:,1],\n",
    "                                       'targ_x_age':clf_x_age.predict(X_x_age),\n",
    "                                       'targ_x': clf_x.predict(X_x)})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "      <th>x</th>\n",
       "      <th>eta</th>\n",
       "      <th>sim_lin</th>\n",
       "      <th>sim_prob</th>\n",
       "      <th>targ_status</th>\n",
       "      <th>prob_x_age</th>\n",
       "      <th>prob_x</th>\n",
       "      <th>targ_x_age</th>\n",
       "      <th>targ_x</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>1.624345</td>\n",
       "      <td>-0.122474</td>\n",
       "      <td>1.112098</td>\n",
       "      <td>0.752520</td>\n",
       "      <td>1</td>\n",
       "      <td>0.775692</td>\n",
       "      <td>0.571468</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>-0.611756</td>\n",
       "      <td>0.228170</td>\n",
       "      <td>-1.088939</td>\n",
       "      <td>0.251818</td>\n",
       "      <td>0</td>\n",
       "      <td>0.242198</td>\n",
       "      <td>0.237208</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>52.0</td>\n",
       "      <td>-0.528172</td>\n",
       "      <td>-0.352305</td>\n",
       "      <td>-1.363402</td>\n",
       "      <td>0.203688</td>\n",
       "      <td>0</td>\n",
       "      <td>0.205344</td>\n",
       "      <td>0.247195</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>20.0</td>\n",
       "      <td>-1.072969</td>\n",
       "      <td>-0.830553</td>\n",
       "      <td>-5.156024</td>\n",
       "      <td>0.005732</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005765</td>\n",
       "      <td>0.187196</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>69.0</td>\n",
       "      <td>0.865408</td>\n",
       "      <td>-0.261090</td>\n",
       "      <td>1.739299</td>\n",
       "      <td>0.850598</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862929</td>\n",
       "      <td>0.448610</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth  age_yr         x       eta   sim_lin  \\\n",
       "0  WHITE ALONE, NOT HISPANIC OR LATINO    55.0  1.624345 -0.122474  1.112098   \n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO    55.0 -0.611756  0.228170 -1.088939   \n",
       "2  WHITE ALONE, NOT HISPANIC OR LATINO    52.0 -0.528172 -0.352305 -1.363402   \n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO    20.0 -1.072969 -0.830553 -5.156024   \n",
       "4  WHITE ALONE, NOT HISPANIC OR LATINO    69.0  0.865408 -0.261090  1.739299   \n",
       "\n",
       "   sim_prob  targ_status  prob_x_age    prob_x  targ_x_age  targ_x  \n",
       "0  0.752520            1    0.775692  0.571468           1       1  \n",
       "1  0.251818            0    0.242198  0.237208           0       0  \n",
       "2  0.203688            0    0.205344  0.247195           0       0  \n",
       "3  0.005732            0    0.005765  0.187196           0       0  \n",
       "4  0.850598            1    0.862929  0.448610           1       0  "
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Concatenate data, predictions to simplify future steps\n",
    "pop_pred_data = pd.concat([pop_data, prediction_df], axis=1)\n",
    "pop_pred_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "           prob_x_age          \n",
      "                  min       max\n",
      "targ_x_age                     \n",
      "0            0.000500  0.499857\n",
      "1            0.500211  0.998124\n",
      "          prob_x          \n",
      "             min       max\n",
      "targ_x                    \n",
      "0       0.041076  0.499923\n",
      "1       0.500050  0.864371\n"
     ]
    }
   ],
   "source": [
    "# Show status boundaries - verify decision thresholds near 50%\n",
    "print(pop_pred_data[['targ_x_age', 'prob_x_age']].groupby('targ_x_age').agg(['min', 'max']))\n",
    "print(pop_pred_data[['targ_x', 'prob_x']].groupby('targ_x').agg(['min', 'max']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Confusion matrices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HISPANIC OR LATINO</th>\n",
       "      <th colspan=\"2\" halign=\"left\">WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>targ_x_age</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>targ_status</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3368</td>\n",
       "      <td>250</td>\n",
       "      <td>2607</td>\n",
       "      <td>453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>563</td>\n",
       "      <td>819</td>\n",
       "      <td>500</td>\n",
       "      <td>1440</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race_eth    HISPANIC OR LATINO      WHITE ALONE, NOT HISPANIC OR LATINO      \n",
       "targ_x_age                   0    1                                   0     1\n",
       "targ_status                                                                  \n",
       "0                         3368  250                                2607   453\n",
       "1                          563  819                                 500  1440"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(pop_pred_data['targ_status'], [pop_pred_data['race_eth'], pop_pred_data['targ_x_age']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HISPANIC OR LATINO</th>\n",
       "      <th colspan=\"2\" halign=\"left\">WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>targ_x</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>targ_status</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3344</td>\n",
       "      <td>274</td>\n",
       "      <td>2828</td>\n",
       "      <td>232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1072</td>\n",
       "      <td>310</td>\n",
       "      <td>1570</td>\n",
       "      <td>370</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race_eth    HISPANIC OR LATINO      WHITE ALONE, NOT HISPANIC OR LATINO     \n",
       "targ_x                       0    1                                   0    1\n",
       "targ_status                                                                 \n",
       "0                         3344  274                                2828  232\n",
       "1                         1072  310                                1570  370"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(pop_pred_data['targ_status'], [pop_pred_data['race_eth'], pop_pred_data['targ_x']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model probability distributions by race/ethnicity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to return KDE estimates\n",
    "def kde_est_race_eth(X, selected_race_ethnicity, data, prob_col = 'prob_x_age'):\n",
    "    \"\"\" Return the Gaussian kernel density estimates for probability distributions.\n",
    "    Selects one race/ethnicity groups and then fits the KDE.  Returns a data frame\n",
    "    containing the distribution.\n",
    "    Returns a data frame containing both distributions.\n",
    "        Input parameters:\n",
    "            X: The x-values for the probability distribution.  Probably array of values in [0,1]\n",
    "            selected_race_ethnicity: String containing the racial/ethnic group to analyze\n",
    "            data: Data for the population of interest, with predictions (model inputs and outputs)\n",
    "            prob_col: Name of the column containing the probabilty for which to generate a KDE \"\"\"\n",
    "    sel_pred = data[data['race_eth'] == selected_race_ethnicity]\n",
    "    sel_gauss_kde = gaussian_kde(sel_pred[prob_col].astype('float'))\n",
    "    score_dist = sel_gauss_kde(X)\n",
    "    score_dist_df = pd.DataFrame.from_dict({'x': X,\n",
    "                                            'race_eth': selected_race_ethnicity,\n",
    "                                            'dist': score_dist})\n",
    "    return score_dist_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>x</th>\n",
       "      <th>race_eth</th>\n",
       "      <th>dist_x_age</th>\n",
       "      <th>dist_x</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>1.541730</td>\n",
       "      <td>0.002089</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.005025</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>1.614963</td>\n",
       "      <td>0.003562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.010050</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>1.682944</td>\n",
       "      <td>0.005877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.015075</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>1.744934</td>\n",
       "      <td>0.009395</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.020101</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>1.800295</td>\n",
       "      <td>0.014560</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          x                             race_eth  dist_x_age    dist_x\n",
       "0  0.000000  WHITE ALONE, NOT HISPANIC OR LATINO    1.541730  0.002089\n",
       "1  0.005025  WHITE ALONE, NOT HISPANIC OR LATINO    1.614963  0.003562\n",
       "2  0.010050  WHITE ALONE, NOT HISPANIC OR LATINO    1.682944  0.005877\n",
       "3  0.015075  WHITE ALONE, NOT HISPANIC OR LATINO    1.744934  0.009395\n",
       "4  0.020101  WHITE ALONE, NOT HISPANIC OR LATINO    1.800295  0.014560"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Generate KDEs for both models\n",
    "score_dist_x = np.linspace(0, 1, 200)\n",
    "kde_x_age = pd.concat([kde_est_race_eth(score_dist_x, 'WHITE ALONE, NOT HISPANIC OR LATINO', pop_pred_data),\n",
    "                          kde_est_race_eth(score_dist_x, 'HISPANIC OR LATINO', pop_pred_data)], axis=0) \\\n",
    "    .rename(columns = {'dist':'dist_x_age'})\n",
    "kde_x = pd.concat([kde_est_race_eth(score_dist_x, 'WHITE ALONE, NOT HISPANIC OR LATINO', \n",
    "                                    pop_pred_data, 'prob_x'),\n",
    "                          kde_est_race_eth(score_dist_x, 'HISPANIC OR LATINO', \n",
    "                                           pop_pred_data, 'prob_x')], axis=0) \\\n",
    "    .rename(columns = {'dist':'dist_x'})\n",
    "kde_df = kde_x_age.merge(kde_x, on=['x', 'race_eth'])\n",
    "kde_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to plot and save \n",
    "def dist_plot_save(data, value_col = 'dist_x_age',\n",
    "                   plt_title = 'Logistic regression model for y=x=age',\n",
    "                   x_label = 'model probability',\n",
    "                   out_fn = None):\n",
    "    \"\"\" Plots the KDE from one of the models.  Optionally saves the plot \"\"\"\n",
    "    maxv = np.max(data[value_col]) * 1.05\n",
    "    pivot_data = data.pivot(index='x', columns='race_eth', values=value_col)\n",
    "    labels = [ '\\n'.join(wrap(l, 20)) for l in pivot_data.columns]\n",
    "    \n",
    "    plt.figure()\n",
    "    pivot_data.plot()\n",
    "    plt.vlines(0.5, 0, maxv, linestyles='dashed')\n",
    "    plt.gca().set_ylim(0, maxv)\n",
    "    plt.legend(title=None, framealpha=1, labels=labels,\n",
    "              loc='upper right')\n",
    "    plt.xlabel(x_label)\n",
    "    plt.tick_params(axis='y', which='both', left=False, labelleft=False)\n",
    "    plt.gca().xaxis.set_major_formatter(matplotlib.ticker.PercentFormatter(1))\n",
    "    plt.ylabel('distribution')\n",
    "    plt.title(plt_title)\n",
    "    if (out_fn is not None):\n",
    "        plt.savefig(out_fn, bbox_inches='tight')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "# Plot the x_age model distributions\n",
    "dist_plot_save(kde_df, value_col = 'dist_x_age',\n",
    "               plt_title = 'Logistic regression model for y=x=age',\n",
    "               out_fn = os.path.join(local_save_path, 'PLOT_model_prob_dist_x_age.png'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEWCAYAAACDoeeyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdeXxM1//H8dfJZF+EELGFrCKIxL5vpbqhVVq6qK2bb/sr1X670PVbX75taXWjaEttsbX2rfYiitDYBdkQhEgiJCHb+f1xR5oQJGTmTibn+XjMQ+beO/e+J4lP7px77jlCSomiKIpinWz0DqAoiqKYjiryiqIoVkwVeUVRFCumiryiKIoVU0VeURTFiqkiryiKYsVUkS+nhBA/CiE+vIfX1RVCXBVCGEyRy5IIIdYIIQbpnaM0hBCDhRDbS7jtTCHE2DusHyuESBZCnC+7hEp5Y6t3gIpACBEPvCil3FBW+5RSvnovx5ZSngJcyyqHJZNSPqJ3Br0IIeoCbwH1pJQX9M6j6EedySv3RAhRpicIFeGThZnVBS7dS4Ev65+toi9V5HUkhHAQQkwSQpw1PiYJIRwKrX9HCHHOuO5FIYQUQgQY1xV8VBdCVBNCrBRCpAkhUoQQ24QQNkKI2Wj/2VcYm2jeEUL4GPdja3ythxBihvEYqUKIpbfJOlgIsUMI8bUQ4hLwiTH/BCHEKSFEkrEJyakU+acIIVYLITKArkKIWkKI34QQF4UQcUKINwrtq5UQIlIIkW481lfG5Y5CiDlCiEvG979HCOFlXLdFCPGi8WsbIcQHQogEIcQFIcQsIYS7cd2N78kg43tJFkKMucPPbaYQYrKxOeiq8ftSw/jzSxVCHBNCNC20fbAxS5oQ4rAQonehdVWFEMuN72s34H/TsRoIIdYbf67RQoin7/hLpb2mO7AeqGXMN9O4vLfx+GnGPMGFXhMvhHhXCHEAyLi50AshfhBCTLxp2XIhxJt3y1Noe3/j+2hmfF7L+LPuUtJ9KPdASqkeJn4A8UD3Ypb/B/gLqA54AhHAZ8Z1DwPngUaAMzAHkECAcf1MYKzx6/HAj4Cd8dEREMUdG/Ax7sfW+HwVsACoYnxt59u8h8FALvB/aM18TsDXwHLAA3ADVgDjS5H/MtAe7WTDGdgLfATYA35ALPCQcfudwEDj165AG+PXrxiP6wwYgOZAJeO6LWhNVQBDgZPG/boCvwOzb/qeTDe+r1DgOhB8m+/FTCDZeCxHYBMQB7xgzDAW2Gzc1s543NHG9/UAcAUIMq6fDywEXIDGQCKw3bjOBTgNDDF+z5saj9vw5t+BYjJ2Ac4Uel4fyAAeNGZ6x5jLvtDvSRTgDTgVs79WwFnAxvi8GpAJeBmfrwTSbvNYWWg/LwFHjD+vdcAEvf9/WvtD9wAV4cHti3wM8Gih5w8B8cavf8FYMI3PA7h9kf8PsOzGujsdu1BBswVqAvlAlRK8h8HAqULPhbFo+Bda1haIK0X+WYXWty68f+Oy94EZxq//BD4Fqt20zVC0P45Nism8hX+K/EbgX4XWBQE5xu/Dje9JnULrdwMDbvO9mAlML/T8/4CjhZ6HAGnGrzui/bGzKbQ+HPgE7Q9CDtCg0Lpx/FPk+wPbbjr2VODjm38HisnYhaJF/kNgYaHnNmh/ULoU+j0ZepffgaPAg8avXwdW3+P/h+XAQeAA4GDq/38V/aGaa/RVC0go9DzBuOzGutOF1hX++mZfop2V/SGEiBVCvFfC43sDKVLK1BJuXziDJ8azb+PH/zRgrXE5lCx/4WX10JoX0grtbzTgZVw/DO1s9JixSaancflstDPC+cZmoS+EEHbFHKu477Vtof2DVoxvyOTOF6iTCn2dVczzG6+tBZyWUubfdOzaaN8rW4p+HwpnrAe0vul78hxQ4w65bqfI+zfmOW3MccOdfscAfgWeN379PNr3/l5MR/vU8p2U8vo97kMpIVXk9XUW7T/yDXWNywDOAXUKrfO+3U6klFeklG9JKf2A3sAoIUS3G6vvcPzTgIcQonIJ8xbeVzJaMWskpaxsfLhLKW8Ut5LkL7y/02ifAioXerhJKR81vscTUspn0Jq2PgcWCyFcpJQ5UspPpZQNgXZAT7Rmk5sV973OpWhxNoWzgLcQovD/tbpoZ9EXjRm8b1p3w2lg603fE1cp5fB7zFHw/oUQwnjcxELb3G1I2jnA40KIUCAYKLh+U+j6RHGPNYW2cwUmAT+jXdfxuIf3opSCKvLmY2e8SHjjYYv2sf0DIYSnEKIaWnv0HOP2C4Ehxot2zmgft4slhOgphAgw/se9DOShNcOAVsT8inudlPIcsAaYLISoIoSwE0J0KsmbMZ4JTge+FkJUN+aoLYR4qLT5jXYDV4wX/5yEEAYhRGMhREvjvp8XQngaj5tmfE2+EKKrECJEaL1z0tGaP/KL2X848KYQwtdYaMYBC6SUuSV5v/dhF9qngneM398uQC9gvpQyD+3awCdCCGchREOgcL/+lUB9IcRA42vthBAtC18wLYWFwGNCiG7GTzpvoV13iCjpDqSUZ4A9aGfwv0kpswqte8T4B6i4R+GurN8AkVLKF9GuB/14D+9FKQVV5M1nNdqZ743HJ2gX6CLR2iYPAvuMy5BSrgG+BTajNcX8ZdxPcR9vA4ENwFW0C5STpZSbjevGo/0hSRNCvF3MaweiFcZjwAVgZCne07s3sgkh0o0Zgu4hP8aC1xMIQ7uImQz8BLgbN3kYOCyEuIpWKAYYi0wNYDFagT8KbKX4ZoRfjMv/NO7/GlpbuklJKbPRivojaO9pMvCClPKYcZPX0Zp2zqO1sc8o9NorQA9gANqZ+Hm0TzEFPbBKkSMarYnlO2OOXkAvY77S+BXtmkOpm2qEEI+j/RxvfBIZBTQTQjxX2n0pJXejB4Zi4Yxnb4fQLlSZ+uyzzJX3/IrG+ElvDtpNVqp4lAPqTN6CCSH6CK0vehW0M7gV5alAlvf8SlHGZp4RwE+qwJcfqshbtlfQmlBi0NrZ7+WCm57Ke37FyPhJLA2t2+0kneMopaCaaxRFUayYOpNXFEWxYiYZiKhatWrSx8fHFLtWlPsSHR0NQFBQkM5JFOVWe/fuTZZSet59y5IzSZH38fEhMjLSFLtWlPvSpUsXALZs2aJrDkUpjhAi4e5blY5qrlEURbFiatxopUL54IMP9I6gKGalirxSoXTv3l3vCIpiVqrIKxVKVFQUAGFhYSbZf3Z2NjExMWRmZppk/4p1cHZ2xt/fH3t7e5MfSxV5pUIZOVIbmsdUF15jYmKoXLkyQUFB2NioS17KrfLz80lKSiImJobg4HsZa6501G+hopShzMxMvLy8VIFXbsvGxgYvLy+zfdpTv4mKUsZUgVfuxpy/I+q3sQLIz5fk5avhKxSlIlJF3spIKTlyNp2v/ohmwLSdhH76B36jVxP0wRoemLiFj5Yd4ui5dL1jKibk6lp01sKZM2fy+uuvA/DJJ58wYcIEAP766y9at25NWFgYwcHBfPLJJwXbe3p6EhYWRsOGDZk+fXrBvpYuXYoQgmPHjhUsi4+PRwjBd999V7Ds9ddfZ+bMmQAMHjyYxYsXA5CTk8N7771HYGAgzZo1o23btqxZUzBxVIHs7GxGjhxJQEAAgYGBPP7445w5c6ZgvcFgICwsjMaNG9OrVy/S0tJu2Udx34vCRo4cSe3atcnPz+fgwYOEhYURFhaGh4cHvr6+hIWF0b17d+Lj42ncuDGgXcsRQrBixYqC/fTs2bPgGs/dcutBXXi1AlJKjpxLZ/XBc6w+eJ645AxsBDSu7c5jTWpSzdWBnLx8TiRdZf6e08zamcBTzevwQc+GuDsVNx2q9Ro3bpzeESzGoEGDWLhwIaGhoeTl5RUM+QDQv39/vv/+ey5cuECjRo3o3bs3Xl5ehIeH06FDB8LDw/n0008Ltq9evTrffPMNr7zyyh17jHz44YecO3eOQ4cO4eDgQFJSElu3br1lu9GjR3PlyhWio6MxGAzMmDGDJ598kl27diGEwMnJqaCn1KBBg/jhhx8YM2ZMid97fn4+S5Yswdvbm61bt9K1a9eC/Q0ePJiePXvSr18/QPsjVlidOnX473//S69evUqdWw+qyJdDUkrikjOIjE9ld3wKu+IucTolC4ONoK1fVV7q6MdDjbyo6nrrBEKpGdlM/TOWaX/GEBFziVnDWuHveaf5qq1Lu3bt9I5gMS5cuEDNmjUB7cy4YcOGt2xTvXp1/P39SUhIwMXFhe3bt7N582Z69epVpMh7enrSvn17fv31V1566aVij5eZmcn06dOJi4vDwUH73fTy8uLpp5++ZbsZM2YQFxeHwWAAYMiQIfzyyy9s2rSJbt26Fdm+bdu2HDhwoFTvfcuWLTRq1Ij+/fsTHh5O165dS/za0NBQcnJyWL9+PQ8++OA95zYXVeTLASklx85fISLmEnviUohMSCH5qjZrm4eLPc3rVWF454DbFvbCqrjY894jDXiokRcvzYrkqR93MmtoKxrXdr/j66xFRIQ2pak5iv2nKw5z5GzZNo01rFWJj3s1uuM2WVlZRe4DSElJoXfv3rds9+abbxIUFESXLl14+OGHGTRoEI6OjkW2iY2NJTY2loCAAJYtW8bDDz9M/fr1qVq1Knv37qV58+YF27777rs88sgjDB06tNhcJ0+epG7dulSqVOmO+W+3XYsWLTh8+HCRYpmXl8fGjRsZNmzYHfd5s/DwcJ555hkef/xxRo8eTU5ODnZ2Jf9UO2bMGD788MMiRb40uc1JFXkLln4thzl/JTBv1ynOpGpzJnt7ONGpvictfTxo6eOBv6fLPx8DpYSECDi8FOK3w5WzkJsNVepBrabQuC/4dQEbA03rVmHRq+14/qddDJ6xmyX/ao+3h7Nu79VcRo8eDVj3AGWFmzJAa2MvbsDAjz76iOeee44//viDefPmER4eXvB9WbBgAdu3b8fBwYGpU6fi4eFBeHg4I0aMAGDAgAGEh4cXKfJ+fn60bt2aefPmmfYN8s8fssTERIKDg4sU27vJzs5m9erVfPXVV7i5udG6dWvWrVtHz549S7yPTp20+e63b99e6uzmpoq8hVp76Bzv/X6QtMwcOgRU4/WuAXQO8qSmu1PxL4jdChs+gbP7wNYJ6rXVHgZ7SImDoyshai54BkOPzyDwQXyrufDr0Jb0mRzBsF/38Nvwdrg5Vqw2elO62xm3JfD392f48OG89NJLeHp6cunSJeCfNvkbUlJS2LRpEwcPHkQIQV5eHkIIvvzyyyL7Gz16NP369aNz5863HCsgIIBTp06Rnp5+x7N5f39/Tp06xZUrV3BzcytYvnfv3oJCfOMPWWZmJg899BA//PADb7zxRone87p160hLSyMkJATQmlmcnJxKVeRBO5sfO3Ystra2Jc6tB9W7xsJIKfls5RFenbOPeh7OrHi9A3NebM2AVnWLL/BZafDbizCrN2Qmw2NfwTuxMHAJPDYRHh4Pz86Ht4/Dkz9B7jWY2w+WDIfrVwmo7saPzzcn5mIGHyw9hJoprOJYtWpVwc/7xIkTGAwGKleuXOy2ixcvZuDAgSQkJBAfH8/p06fx9fVl27ZtRbZr0KABDRs2LNL75AZnZ2eGDRvGiBEjyM7WmhsvXrzIokWLimzn4uLCoEGDGDVqFHl5eQDMmjWLzMxMHnjggVv2+e233zJx4kRyc0s2fXB4eDg//fQT8fHxxMfHExcXx/r160t9c1KPHj1ITU0tuB5QmtzmpIq8BZFSMm71UX7eHsegtvVY9Go7Qurcoa08cR9M7QSHl0Dnd+G13dByGNgX0+xi5whNntK26fwuHJgP07pAShztA6oxolsgy6LOsuTvRJO9P8WyzJ49m6CgIMLCwhg4cCBz584tuGB4s/DwcPr06VNkWd++fQkPD79l2zFjxty22+DYsWPx9PSkYcOGNG7cmJ49exZ7Vj9+/HgcHR2pX78+gYGBLFq0iCVLlhTbQ6Vp06Y0adKk2CyZmZnUqVOn4DFu3DjWrl3LY489VrCNi4sLHTp0KPYP092MGTOG06dP31NuczHJHK8tWrSQatKQ0vt5exyfrTzCoLb1+KR3o9v/YkgJu6fBujHg6gVPzQDvVqU7WNw2WDgQbGzh2YXk1WzKM9P+4vDZy6wf1ZlalW/TLFTOmXrSkJsvRirK7RT3uyKE2CulbFGWx1Fn8hbiyNl0Pl9zjAcbevFxrzsU+NzrsORVWPMO+D8Ar24rfYEH8O0Iw9aDnRPMegLDub+Z+HQo+RI+tOJmm0mTJjFp0iS9YyiK2agibwGu5eQxYv7fVHa24/O+TbCxuU2Bz0iGWY9rTS1dRsMz88HZ494PXC0QBq8GJ3eY/QTe2bG81aM+G49dYNXBc/e+Xwt2465GRakoVJG3ADN2xHPiwlU+79cED5fb3C144RhMfwDO/g39foEu70JZDHJU2RsGrQQ7F5jbj8GNbGlcuxJjVx4lM7tkF7LKkw0bNrBhwwa9YyiK2agir7Pkq9f5YfNJugdXp2tQ9eI3OrkRfn4QcrJg8Cqtv3tZqlIPnl8M2RnYzu/Ppw/7cD79GlO3xpbtcSzA2LFjGTt2rN4xFMVsVJHX2dfrj3MtJ4/3H73N5AH7ZsPcp6ByXXhpE9Qp02sy//BqBE/PgovHaB71IY81rsHUP2M4dznLNMdTFMUsVJHX0dm0LBbsOc2AVt7Fjx8T8R0sf127S3XoWq1pxZT8u0K3j+DwEsbW3Ea+hC/WRt/9dYqiWCxV5HU0fZvWHPJqZ/+iK6SETWPhjw+g4RPaBVYHt2L2YALtR0KDnlTZ/h8+bpLGkr8TiTpd/DCuiuV58803i/Qeeuihh3jxxRcLnr/11lt89dVXRYbPvaHwMMQ3hgfu06cPYWFhBAQE4O7uXnDhOiIigi5duhT0sw8LCysYtbE4TzzxBG3atLnt8QpbunQpTZo0ITg4mJCQEJYuXVqwbvDgwdSuXZvr168DkJycjI+PD6CNFunk5FSQJywsjFmzZt3x+3Wn/QEcPnyYBx54gKCgIAIDA/nss8+QUjJjxoyCY9jb2xMSEkJYWBjvvffeHY+nBzWsgU5SMrKZv/s0vcNqUadKoZuXpIR1o+GvydB0IPT6BmyKv0HFJISAJybDtK48k/ARs1zG8dnKIyx+ta2uN3QoJdO+fXsWLlzIyJEjyc/PJzk5mfT0fwZJi4iI4Ouvvy7x/pYsWQJo9xVMmDCBlStXFlk/d+5cWrS4cxNiWloae/fuxdXVldjYWPz8/G677f79+3n77bdZv349vr6+xMXF8eCDD+Ln50eTJk0AbcTMX375heHDh9/yen9//yLj9pTE7faXlZVF7969mTJlCj169CAzM5O+ffsyefJkXnvtNYYMGQKAj48Pmzdvplq1aqU6rrmoM3mdzNoZT1ZOHsNvPov/80utwLceDr2/M2+Bv8HRHfrPweb6FWZ4zGBvQgobj14wfw4TmDp1KlOnTtU7hsm0a9eOnTt3AtpZaOPGjXFzcyM1NZXr169z9OhRmjVrZtZMv//+O7169WLAgAHMnz//jttOmDCB0aNH4+vrC4Cvry/vv/9+kTFyRo4cyddff13iYQzu5nb7mzdvHu3bt6dHjx6ANoTC999/z//+978yOa65qCKvg5y8fObtOkWXIE8CvQo1w/w9Bzb/F0KfgYfGaWfVevFqCD0+o9bFHYystJWJ64+TbwVTCAYFBREUFKR3DJOpVasWtra2nDp1ioiICNq2bUvr1q3ZuXMnkZGRhISEFEzqERMTU6Rp48cffyz18Z577rmC1//73/8udpsbw/o+88wzxQ49UNjhw4dvuQv0xlC9N9StW5cOHTowe/bsW15/83u6eWyd4txuf8Vl8ff35+rVq0U+HVk61Vyjgz8OJ3HhynX+17bePwsT98HKUeDb2XgGbwF/f1u+CMfX8X+xv7LifCBrDgXwWJOaeqe6LzfGJyluVh9r0a5dOyIiIoiIiGDUqFEkJiYSERGBu7s77du3L9ju5qaNG9P/lcbdmmuSkpI4ceIEHTp0QAiBnZ0dhw4duuV6QGm9//77PP7440XGoIF7a6650/6sgQVUkopn9l/x1KniROf6xn7xWamw8AVwrQ79ZoDBQob7FQIe/x4bBxemOP3Id+sPl/sJwSdOnMjEiRP1jmFS7du3JyIigoMHD9K4cWPatGnDzp07iYiIMPvMWAsXLiQ1NRVfX198fHyIj4+/49l8w4YN2bt3b5Fle/fupVGjosM2BwYGEhYWxsKFC8skZ3H7Ky5LbGwsrq6ud534xJKoIm9mJy9c4a/YFJ5rXQ/DjeEL1o6G9LPw1K/gUlXfgDdzq4Ho/S3182PomTqL5fvVKJWWrl27dqxcuRIPDw8MBgMeHh6kpaWxc+dOsxf58PBw1q5dWzCs7969e+/YLv/2228zfvz4gnlV4+PjGTduHG+99dYt244ZM6bY3jnFeeGFF9i9e/cdt7l5f8899xzbt28vuEM6KyuLN954g3feeadEx7QUqsib2cLIM9jaCPo1r6MtOL4O9s+DjqOgjoWOXhjcCxn2HMNtV7Bu3Spy8vL1TqTcQUhICMnJyUW6LIaEhODu7l7mPUAKt8l37969yLr4+HgSEhKK5PD19cXd3Z1du3YB2h3IhYcCDgsL4/PPP6dXr140aNCAXr168cUXXxQ73lCjRo1uuYh8c5v8t99+C8CBAweoVavWHd/LzftzcnJi2bJljB07lqCgIEJCQmjZsiWvv/566b5JOlNDDZtRTl4+bcdvolndykx7oQVkZ8L3LcGxEry8BWzvPD+rrq6lk/Vta85elfz96DL6tSmfFy/VUMMVT3p6OsOGDbtlchK9qaGGrdCW6IskX73OUy2Md67u/B7Sz2gzOFlygQdwrIRj3yn425yDDf8p923zSsVRqVIliyvw5qR615jRwsjTVHN1oEuQJ6Sfg+1fQ3BvqGfedtJ7Jfy7EO//HP1i5rJj01Lad+9z9xdZmOK63ZnMmvfg/MGy3WeNEHikfPXTVvSlzuTNJDUjm83HLtCnaS3sDDaw9X+QlwMPfqp3tFKp+/SXnLGpRcCOd8jPuqx3nFLz9vbG29vEYwApiiWRUpb5o3nz5lIpas5f8bLeuyvlwTNpUqaekvLTqlKueFPvWPdk8/rlMvcjd3l65jC9o5Ta/Pnz5fz58022/8jISJPtu6RcXFyKPJ8xY4Z87bXXpJRSfvzxx/LLL7+UUkq5c+dO2apVKxkaGiobNGggP/7444Ltq1WrJkNDQ2VwcLCcNm1awb6WLFkiAXn06NGCZXFxcRKQ3377bcGy1157Tc6YMUNKKeWgQYPkokWLpJRSZmdny3fffVcGBATIpk2byjZt2sjVq1ff8h6uX78uR4wYIf39/WVAQIDs3bu3PH36dMF6GxsbGRoaKhs1aiR79uwpU1NTS/S9KGzEiBGyVq1aMi8vTx44cECGhobK0NBQWaVKFenj4yNDQ0Nlt27dZFxcnGzUqJGUUsrNmzdLQC5fvrxgP4899pjcvHlziXIXVtzvChApy7geqzN5M1n291kCqrvSqFYl2GEcQKrDm/qGukcduj5GuN0T1IlbhIxeq3ecUpkyZQpTpkzRO4ZFGDRoENOmTSMqKopDhw7x9NNPF6zr378/UVFRbNmyhdGjR5OUlARoXSI7dOhwS1/36tWr880335CdnX3HY3744YecO3eOQ4cOsW/fPpYuXcqVK1du2W706NFcuXKF6OhoTpw4wRNPPMGTTz5ZMC2lk5NTQW4PDw9++OGHUr33/Px8lixZgre3N1u3biUkJISoqCiioqLo3bs3X375JVFRUcVOMFOnTh3++9//Frvfu+XWgyryZpCYlsXu+BQeD62FuHIe9s2CsGdNP3SwidgabDA8MIZj+d7kLHkdMlP0jqTcgwsXLlCzpnYHs8FgoGHDhrdsU716dfz9/UlISODq1ats376dn3/++Za+7p6ennTr1o1ff/31tsfLzMxk+vTpfPfddzg4aB0NvLy8ivxxubHdjBkz+PrrrzEYtLGbhgwZgoODA5s2bbplv23btiUxsXT3b2zZsoVGjRoxfPjwuw61cLPQ0FDc3d1Zv379feU2F1XkzWDF/rMA9A6rBXt+0triO4zUOdX96dPSj7H2b2BzLQVWFz9miaKPrKysIn3FP/roo2K3e/PNNwkKCqJPnz5MnTqVa9eu3bJNbGwssbGxBAQEsGzZMh5++GHq169P1apVb7kb9N1332XChAnk5eUVe7yTJ09St27du94tervtbh7DBiAvL4+NGzfSu3fvO+7zZjfG0+nTpw+rVq0iJyenVK8fM2bMLTOMlSa3OakibwZL/06kad3K1KtkA5G/QNCj4HH74VbLA0c7Ax07duObnD5waDEc+l3vSIrRjaaMG4///Oc/xW730UcfERkZSY8ePZg3bx4PP/xwwboFCxYQFhbGM888w9SpU/Hw8CA8PJwBAwYAMGDAgFvOgP38/GjdujXz5s0z3ZszuvGHrEaNGiQlJfHggw+W+LXZ2dmsXr2aJ554gkqVKtG6dWvWrVtXquN36tQJgO3bt5fqdXpQRd7Eos9f4dj5KzweWgsOLoKsFGhz6zjY5dGAVnX51fAkCY7BsHIkXD6jdySllPz9/Rk+fDgbN25k//79XLp0CfinTX7Xrl306dOHlJQUNm3axIsvvoiPjw9ffvklCxcuvKWtefTo0Xz++efFtkEHBARw6tSpu47g6O/vz6lTp25pqy88hs2NP2QJCQlIKUvVJr9u3TrS0tIICQnBx8eH7du3l7rJBm49my9Jbj2oIm9iy/cnYrARPBZSE/76Ebwag08HvWOVCXcnO/q19GVo+svk5+XAklch37KHPFi8eDGLFy/WO4ZFWLVqVUExPnHiBAaDgcqVKxe77eLFixk4cCAJCQnEx8dz+vRpfH19bxnKt0GDBjRs2LBgtM/CnJ2dGTZsGCNGjCi4QHvx4sVbblRycXFh0KBBjBo1qqDpZ9asWWRmZvLAAw/css9vv/2WiRMnlnh8+fDwcH766aeC8XTi4uJYv349mZmZJXr9DT169CA1NZUDBw6UOmucQNgAACAASURBVLc5qSJvQlJKlkWdpX1ANTwvH4ALh7Xhe61ohqUh7X2Ik16sqTMS4rfBzu/0jnRH1apVs9gZfMxt9uzZBdP3DRw4kLlz5xZcMLxZeHg4ffoUvfmtb9++xZ4BjxkzhjNniv9UN3bsWDw9PWnYsCGNGzemZ8+exbbRjx8/HkdHR+rXr09gYCCLFi1iyZIlxc5O1rRpU5o0aVJslszMzCJj44wbN461a9cWGVLYxcWFDh06FPuH6W7GjBnD6dOn7ym3uaixa0xob0IqfadEMOGpUPqdHg+Hl8Db0eabr9VMXpu3jz+PX+Dv+rOwPfkHvLQRaobqHatYM2fOBLS5PU1BjV2jlJQau8YKrDpwDnuDDT38neDQbxDSz+oKPMBLHf24ci2P+TX+Dc5V4beXtMHXLNDMmTMLCr2iVASqyJtIfr5kzaFzdKpfjUonlkBuFjQfrHcskwjzrkxLnypM2Z1K3uOTITka1hffbU9RFPNSRd5Eos6kce7yNR5pXBP+nqtdcK3VVO9YJvNiRz8S07JYkxUMbV6DPdO1sfIroHwLv/is6M+cvyOqyJvImoPnsDMIelS/DGf3aZNzW9EF15t1D/bCp6oz07fFIbt9CNUbwbLX4OpFvaOZlbOzM0lJSarQK7eVn59PUlISzs7OZjmeGmrYBKSUrD54no6BnrhFLwZhgJCn9I5lUgYbwbAOvny47DCRiVm07PsTTOuiFfpnF1j1H7jC/P39iYmJ4ezZs3pHUSyYs7Mz/v7+ZjmWKvImsP/MZRLTshjZzR+2LYCAbuDmpXcsk+vX3JuJ648z/c9YWr7QQhtGee17cGABhA7QOx4Aq1evNun+7e3tCQ4ONukxFKU0VHONCaw5eA5bG8EjrjGQnghN+usdySyc7A0837oe648mkXApA1q9AnVawroxFjOImbOzs9k+JiuKJVBFvoxJKVl18BztA6rhenIZ2LloY9VUEAPb1sPWRjBjRzzY2EDPryErFTYWP36KuU2ePJnJkyfrHUNRzEYV+TJ2KDGdM6lZ9GxUDY4sg6BHwL7inDl6VXKkZ5NaLIo8Tfq1HG26utavwN6ZkKTfSHw3LFy4kIULF+odQ1HMRhX5MrbqRlONS7R2Btu4r96RzG5YB18ysvNYuMd4u3enf4NjJdV3XlF0oIp8GdJ61ZyjrX9VXE8sBwd37aJrBdO4tjutfD2YsSOe3Lx8cPaATu/AyQ0Qo9/kCYpSEakiX4YOn03nVEomvRpWhWMrIbgn2DroHUsXQ9v7kpiWxR9HtGnjaPUSuNeFTf8FHadCU5SKRhX5MrT64DkMNoJHnA7D9XRo9KTekXTzYEMv6no488v2OG2BrQN0HAWJkepsXlHMSBX5MlLQVONXFbeTy8DJA/w66x1LNwYbweB2PkQmpLL/dJq2MOxZqFQHtn6u29n8li1b2LJliy7HVhQ9qCJfRo6eu0L8pUx6BbtD9Bpo2BsMdnrH0tVTLerg6mDLz4XP5juMhNO7IN7yp01TFGuginwZ+ePIeYSARxwOQE5mhexVczM3Rzv6t/Rm9cFznLucpS1s+rw2HPFf+vRVnzBhAhMmTNDl2IqiB1Xky8j6I0k0q1uFSjErwKU61GuvdySLMLidD/lSMmtngrbAzglaDNM+7VyKMXuelStXsnLlSrMfV1H0oop8GUhMy+Lw2XQeCaqkdRMM7gU2xU+jVtF4ezjTo2EN5u06RWa2cQ7Oli9qTVl/TdE3nKJUAKrIl4ENxm6Cj7kc05pqgnvpnMiyDOvoy+WsHH7fl6gtcPOCxv0gah5cv3LnFyuKcl9UkS8DG44m4efpQs2zG8CxMvh00DuSRWlRrwpN6rjzy4448vONvWpaDIWcDDi4WN9wimLlVJG/T+nXcvgr9hIPNagK0au1wcgqeK+amwkhGNrel9iLGWw9bpxEpE4LbWKRvTPNmsXJyQknJyezHlNR9KSK/H3aEn2RnDxJnyqxcO2yaqq5jUdDauJVyYFfdhi7UwqhzXl7LgrORpktx5o1a1izZo3ZjqcoelNF/j6tP5JEVRd7Ai5t0YYV9u+qdySLZG9rwwttfdh2Ipno88Z2+CZPga0j7PtV33CKYsVUkb8P2bn5bIm+QPcGVbE5thICH9S6CCrFerZVXRztbP4Z6sCpCjTqAwcWQXaGWTJ89tlnfPbZZ2Y5lqJYAlXk78PuuBSuXMulr9d5yLigmmruooqLPU82q8OSqEQuXb2uLWw2CLKvwKHfzZJh48aNbNy40SzHUhRLoIr8fdhwNAlHOxuaXt0GBnsI7KF3JIs3tL0P2bn5zN11SltQtw1UCzL7BVhFqShUkb9HUko2HbtAe7+q2B1fCX5dtIkxlDsKqO5G5/qezNqZwPXcvH8uwCZGwvlDesdTFKujivw9ikvO4FRKJk/UToO0U9DgMb0jlRvDOviSfPU6K/ef0xaEDgAbO4iaq28wRbFCqsjfoy3RWn/vjjJSW1D/YR3TlC8dA6sRWN2Vn7fHIaXUZo4KegQOLIDcbJMeu2rVqlStWtWkx1AUS6KK/D3aevwifp4uVD61EWo1A7caekcqN4QQDO3gy5Fz6fwVm6ItbPo8ZF6CE3+Y9Ni//fYbv/32m0mPoSiWRBX5e3AtJ4+/Yi/xmK8tJO7VzkKVUunTtDYeLvb/3Bzl3w1ca6gmG0UpY6rI34OdsZe4nptPT6cDgFRNNffA0c7Ac63rsuFoEvHJGWCwhdD+cHwdXL1gsuO+//77vP/++ybbv6JYGlXk78HW6Is42tkQkLpdm86uRojekcqlgW3qYWsjmHHjbD7sOZB5Wtu8iezcuZOdO3eabP+KYmlUkb8HW6Iv0NHXDUPcZqj/kNYNUCm16pUc6R1am4WRZ0jNyAbPIKjdAv6eq9scsIpibVSRL6X45AziL2XSv2q8Nna8ao+/Ly938iMrJ485fxlnjmr6HFw8Cmf36RtMUayEKvKltCVaay9unbtbG5DMp6POicq3oBpudAny5Ned8VzLydPmxrV11CYUURTlvqkiX0pbj1/Et6ozbgkbtREn7Rz1jlTuvdzJj+Sr2drMUY7u2hhABxdBzrUyP1adOnWoU6dOme9XUSxViYu8EKK2EKKdEKLTjYcpg1miazl57Iy9RH/vNEg/o5pqykhbv6qE1Hbnp22x2sxRYc9pY/NHryrzY82ZM4c5c+aU+X4VxVKVqMgLIT4HdgAfAP82Pt42YS6LtC8hlWs5+XS3268tUAOSlQkhBC938iM2OYMNR5PAtzO4e2sXYBVFuS+2JdzuCSBISnndlGEs3Y6YZAw2At+0XVAzFFyr6x3JajzSuAZ1qjgx7c9YejSqAaHPwJ9fwuVEcK9dZscZOXIkAJMmTSqzfSqKJStpc00sUOEnLt1x8hLtatthSNyt3aGplBlbgw3DOvgSmZDK3oRUCHsWkPB32TatREVFERVlvukGFUVvJS3ymUCUEGKqEOLbGw9TBrM06ddyOHAmjac8YiA/FwK66x3J6jzdwht3Jzum/RkDHr7aH9K9MyAvR+9oilJulbTILwc+AyKAvYUeFcau2BTyJbSWUWDvBt6t9I5kdVwcbBnYph5/HEki9uJVaPkiXDkH0WribUW5VyUq8lLKX4Fw/inu84zLKowdJ5NxtBNUT9oOfp3BUOFbr0xiUDsf7Aw2/LQ9Trub2N0bIn/WO5ailFsl7V3TBTgB/ABMBo5XtC6UO04m07t2BuLyaQhQ7fGm4unmQN9mtVm89wzJmbnQfBDEboHkE2Wy//r161O/fv0y2ZeilAclba6ZCPSQUnaWUnYCHgK+Nl0sy3Ih/RonLlzlcddj2gJ10dWkXuzoR05ePrMi4qHpC9qsUZG/lMm+p02bxrRp08pkX4pSHpS0yNtJKaNvPJFSHqcC9baJiLkEQJNrkVA1EKrU0zmRdfP3dKV7sBez/kogy6EaNOytjTOfnal3NEUpd0pa5COFED8JIboYH9OBSFMGsyQ7Tibj5ZSP6/ldqqnGTF7p5EdaZg4L9pyCFsO0O2AP3f+MTi+//DIvv/xyGSRUlPKhpEV+OHAEeMP4OGJcZvWklOw4mcxzNRMRuVmq66SZtPDxoKVPFab9GUt27TZQvSHs+vG+hyA+fvw4x48fL6OUimL5Stq75rqU8isp5ZPGx9cV5e7X+EuZnL18je52h8DgAPXa6x2pwvhX1wDOXr7Gsv1noe3rkHQIYjbpHUtRypU7FnkhxELjvweFEAdufpgnor4iYpIBCLiyB+q2AXtnnRNVHF3qe9KwZiWmbI0hr3E/cKsJO77RO5ailCt3O5MfYfy3J9CrmIfV2xOXQpBrFvYpx8Cvi95xKhQhBP/q6k/sxQzWHUuBNsMhbiuc/VvvaIpSbtyxyEspzxm//JeUMqHwA/iX6ePpb098Kk9Xi9ee+HbWNUtF9EjjmvhWc2HylpPIZoPAoRLsuPcRNcLCwggLCyvDhIpi2Up64fXBYpZZ/WDqiWlZJKZl0cFwRCsuNUP1jlThGGwEr3b241BiOn+ezoEWQ+DIUkiNv6f9TZo0SY1AqVQod2uTHy6EOAg0uKk9Pg6w+jb5PXEpAPhciQSfDmAo6cjMSlnq07QONSo58sPmk9B6OAgD7PxB71iKUi7c7Ux+Hlrb+zKKtsU3l1I+b+Jsutsdn0J9h1Qc0hPAt0KN4mBR7G1teKmTH7vjUohMcYAm/WHfbMhILvW+nn/+eZ5/3up/dRWlwN3a5C9LKeOBdwFZ6OEqhKhr+nj62hOXwlNV47Qnqsjr6plW3lRxtmPylhhoPwJyr8Ffk0u9nzNnznDmzBkTJFQUy1TSNvlVwErjvxvRJhGx6vFfUzOyOXHhKp1sj4BzNe1mHEU3zva2DGnvy6ZjFzic46UNdbB7OmSl6R1NUSxaSW+GCpFSNjH+Gwi0AnaaNpq+9sSnABLfK3u1s3gh9I5U4Q1q64Obgy3fbzoJHd+C6+mwZ7resRTFopX0TL4IKeU+oHUZZ7Eoe+JTCDIkYZ+VpJpqLIS7sx1D2vuw5tB5jkhfbSL1nZMhO0PvaIpisUo6nvyoQo+3hRDzgLMmzqar3fGp9Ksaoz3xU/3jLcWwDn64Odjy7cYT0PFtyEqBvTNL/Pq2bdvStm1b0wVUFAtT0jN5t0IPB7S2+cdNFUpvmdm5HE68rLXHu3tDFV+9IylG7s52DOngy9rD5zliGww+HSHiO8gt2VBK48ePZ/z48SZOqSiWo6Rt8p9KKT9FmyjkWynlXCnlNdNG00/UqTTy8vPwu/q3ao+3QMPa++LmaMs3G49rbfNXzkHUPL1jKYpFKmlzTQvjTVEHgINCiP1CiBamjaaf3fEpBNucwi47TQ1lYIHcne0Y2t6XdYeTOOzYFGo3hx2TIC/3rq/t27cvffv2NUNKRbEMJW2u+QVt/BofKaUP8JpxmVXaE59CH/eT2hPfjvqGUYo1tIN2Nj9p40mtbT41vkSTily6dIlLly6ZPqCiWIiSFvk8KeW2G0+klNuBu582lUM5efnsS0ijs91Rbaq/SrX0jqQUw93Jjpc7+rH+SBL7nFpr9zFs/wry8/WOpigW5W5j1zQTQjQDtgohphqn/usshJgMbDFLQjM7fDadnJzr+GXuV71qLNzQDr5Uc7Xn87XHkR1GwcVjcGyl3rEUxaLcbcStiTc9/7jQ1/c3D5uF2hOXQhMRi21upuofb+FcHGz5vwcC+Xj5Yf7s3JHOHn6wbQIE91IXyxXF6I5FXkrZ1VxBLMXu+BQedT0OOULrnqdYtGda1WX6tli++OMEHduPxGbFG3ByIwQWPxdvt25qInalYrljkRdCPC+lnCOEGFXceinlV6aJpY/8fElkfArvOh2Fqo3B2UPvSMpd2Nva8FaP+ry5YD+rbTrTs1IdrW3+NkX+ww8/NHNCRdHX3S68uhj/dbvNw6rEXLxKZmYGvlmHVdfJcqR3aG0a1HBjwoY4clu9Cgk74GyU3rEUxSLcrblmqhDCAKRLKb82Uybd7I5PoZnNCQz52arIlyMGG8E7DwcxdGYk4bmdGWjvCrt+hD4/3rLtI49oE5qtWWPVg6gqSoG7dqGUUuYBz5ghi+72xKXQ3eEYUhignhrfpDzpGlSdDgHVmLD1PNcb99f6zF9JumW7rKwssrKydEioKPooaT/5HUKI74UQHW90qzR2rbQqe+JT6epwFFG7OThYXWuUVRNC8EHPYK5cy2HqtR6Qlw2RVnu/nqKUWEmLfBjQCPgPWrfKicAEU4XSQ2JaFmlpKfhcj1ZdJ8upBjUq0b9lXb6NyiejXjeI/LnEA5cpirUqaZEfJqXsWvgBvGjKYOYWGZ9CS5tj2Mg8VeTLsVEP1sfRzsCUaz0g42KJhjpQFGtW0iK/uJhli8oyiN52x6XQxe4o0uAA3q30jqPcI083B17rGsD3CXXIcA/U5oGV/9y317NnT3r27KljQkUxr7v1k2+A1kzjLoR4stCqSoCjKYOZ2574FF6yP4ao3QrsnPSOo9yHIe19mLc7gWnZPXjz8g9waifUawfA22+/rXM6RTGvu53JBwE9gcpAr0KPZsBLpo1mPqkZ2VxIOke9nBjVddIKONoZ+OCxhkxNbcF1WzfY85PekRRFN3frJ78MWCaEaCultNqJuyMTUmltcxSBVO3xVqJHQy9aBNRmwZlODDyyHHElCdy86NKlCwBbtmzRNZ+imEtJ2+T7CCEqCSHshBAbhRAXhRDPmzSZGe2JT6Gj4QjSzgVqW13P0ApJCMHHvRrya3Y3RH4O7PtV70iKoouSFvkeUsp0tKabeCAA+LepQpnb7rgUOtsfRdRrBwY7veMoZSTQy41ObdvwZ34IObt/LtHMUYpibUpa5G9UvseARVLKyybKY3aZ2bkkJcbjnXdaNdVYoZHd6/O77aPYZZxHRq/SO46imF1Ji/wKIcQxoDmwUQjhCVjFRN5Rp9JoyWHtiSryVsfdyY7WPZ7hjKxG8ubJesdRFLMrUZGXUr4HtANaSClzgAzgcVMGM5fd8Sm0Nxwm37Ey1AjRO45iAk+38mGD86N4XvyLIU905emnn9Y7kqKYzd36yT8gpdxUuI+8KDrjzu+mCmYue+JTmGB3FBufDmBj0DuOYgIGG0Fo7//j+vx5NHE8Q9NXP777ixTFStxt+r9OwCa0vvESEDf9W66LfE5ePkkJx6lpSFJNNVauaXB99lTqQoNzKzgRf4pAn7p6R1IUs7hbc80V46xQhwo9DgMHjV+Xa4fPptM0/6D2RBV5q+f76EjcRBZzv35f7yiKYjZ3K/KuaDNANQeGAzWBWsCraHe9lmuR8Sm0szlMnnM18GygdxzFxKo16MDRa9Xo776fiBMX9Y6jKGZxxyIvpfxUSvkpUAdoJqV8W0r5FlrRL/efd3fHXqKT4QgGv85Q9FqDYo2EYHmCC8E2p/lt6WJy8/L1TqQoJlfSLpReQHah59nGZeWWlJLLCVFUJRX8u+kdRzGTTUmVSc+1o3P6MubuOqV3HEUxuZIW+VnAbiHEJ0KIT4BdwExThTKHmItXCb2+V3vi31XfMIrZXM+3Ye25yjxq2MPMP3aRkpF99xcpSjlW0n7y/wWGAKnGxxAp5XhTBjO13XGpdLI5QLZHA6hUS+84ipkMHjwYQ5uXsSWX3rnrmfhHtN6RFMWk7taFsoCUch+wz4RZzCoq9iz9DNHY1X9F7yiKGQ0ePFj7YvZOhp3aQvPdvXmudT0a1qqkay5FMZWSNtdYnbzYbdiTiwhQ7fEVSXJyMsnJydDyJSrlXKS3434+WXEYWWj2KEWxJhWyyJ9Ny6JRViS5Ng4FMwYpFUO/fv3o168f1H8I3L35t8ef7I5LYeWBc3pHUxSTqJBFfk98Ch1tDpJVs7Wa6q+isjFAiyHUuLSbh6pfZvzqo2Rl5+mdSlHKXIUs8tHRRwm0ScQ5+EG9oyh6ajYIDPb8p8Z2zl6+xvRtsXonUpQyVyGLvG38ZgAMgd11TqLoyqUahD2LV8wi+gfZ8uPWGC6kW8UI2opSoMIV+bTMbOpf3cNVe0+oHqx3HEVv7UdCfh6jK68nJy+fr9Yf1zuRopSpEnehtBa7Yy/SweYQWd4P46qGMqhwhg8fXnSBhy80eRr3w3MY3uIJvtt9mkHtfAiuqbpUKtahwp3Jnzm0g8oiA/eQh/WOouigf//+9O/fv+jCjm9BXjav2fxGJUc7xq0+qrpUKlajwhV554SN5GODfX3VHl8RnT59mtOnTxddWC0Qmg/CYf8sPmhjz7YTyWw5rkapVKxDhSry6ddyaJzxF2fdQsDZQ+84ig4GDhzIwIEDb13R5X2wdaRvynR8qjozbtVRNUqlYhUqVJE/eOQojW3iyfXvoXcUxdK4VocOb2ITvZKJYec5ceEqCyJP3/11imLhKlSRv3xwFQA1WljFHORKWWv3BngG0+zgWDrVdeTr9ce5ci1H71SKcl8qVJGvmriFizbVcazdWO8oiiWytYfe3yLSE/m6yiKSr2bz49YYvVMpyn2pMEU+IyODxtf/5oxnRzULlHJ73q2g/QiqRofzme8hftoWx9m0LL1TKco9qzBFPjZyHS7iOrYNVNfJiuytt97irbfeuvNGD3wI9Trw3MWvaEgsX65TY84r5VeFKfLZR9dyTdrh21IV+YqsV69e9OrV684bGWyh3y/YuFRnruMXHIjaze64FPMEVJQyVjGKvJTUvriVgw5huLqqOxkrsujoaKKjS3Bm7uYFLyzFycGeBQ7jmL54Odm5qkulUv5UiCJ//fwxauSd51LNLnpHUXT2yiuv8MorJZwNrKo/4oVluDrZ89XV91i7dJZpwymKCVSIIn8uchkAriGP6ZxEKXe8GuL46mYuO9Sk96ERXF7yb8i9rncqRSmxClHkbU+s5Wh+XUIaqa6Tyj1wr439K5uYJx/Gff805PSucOGo3qkUpUSsv8hnXKJm+n4OuLTD3clO7zRKOVW9ahXyHvmCIdn/5nrqOZjaGSK+h3w1m5Ri2ay+yGcdWY2BfK75q141yv15rlVdsny60ePaeDLrdoE/xsDMxyBFzSilWC6rL/JXopZyVnoQGNZB7yiKBfjggw/44IMP7um1NjaCiU+HkWpThYEZI8jtPQWSjsCU9rB/fhknVZSyYd1FPieLyue2s0W2oLmPGnVSge7du9O9+70PM127shP/7RPC3lNpjEsMg39FQK2msOQVWPMe5KmxbhTLYt1FPnYL9vnXSPTqioOtQe80igWIiooiKirqvvbRO7QWg9v58MuOOJbFCXhhGbQeDrumwOw+cFWNRa9YDqsu8hkHlpMunajauJveURQLMXLkSEaOHHnf+xnzWDCtfD1497cDHE7KhEf+B32mwpk98HN3SE0og7SKcv+st8jn52E4sZYt+WG0rV9T7zSKlbEz2PDDs82o7GTPq3P2kpqRDaEDYPAqyEqFGY9A8km9YyqKFRf5M5E4Zqfwl10bGtRw0zuNYoU83RyY8nwzki5f58VZkWRl50GdFlqhz72uFfqkw3rHVCo4qy3y8uhKcrAl378bQg0trJhI07pVmDQgjH2nUvm/8H3alIE1QmDIGrAxaF0sVaFXdGSdRV5Kso+sZGdeMC0a+OqdRrFyj4bU5D+PN2bD0QuMXnIQKSV41ochq8HWUbsYe0lNPqLow1bvACZx4SgOl2P5I38IrwdU0zuNYkHGjRtnkv0ObFOPi1eu8+3GE3i6OfDvhxqAhx8MXKo128x+Aoaug0q1THJ8Rbkd6zyTP7qcfATHq3Shhruj3mkUC9KuXTvatWtnkn2/2T2QZ1rV5YfNMczcEactrN4Anv8NMlO1M/qMSyY5tqLcjlUW+fzDy9grg2gUFKh3FMXCREREEBERYZJ9CyEY+0RjejT04tOVR1h54Ky2onYzeHY+pMbDvKchO9Mkx1eU4lhfkU8+ic3FI6zObckDDarrnUaxMKNHj2b06NEm27/BRvDtM01pUa8KoxbsJyImWVvh0wH6/gxn98FvwyAv12QZFKUw6yvyR7Wx4/80tKGVrxrKQDE/RzsDP73QknpVnXll1l6OnkvXVgT3hEe+gOjVsPotkFLfoEqFYHVFXh5ZzmERSGBgsBrKQNGNu7Mdvw5thYuDLYNn7CYxLUtb0eol6DAK9s6EPyfomlGpGKyryKcmIM5FsSy7hWqqUXRXq7ITvw5tRWZ2HoN+2c3lLOPgZd0+gtBnYPNY+HuOviEVq2ddRf7ocgDW5LeiSwNPncMoCgTVcGPawBYkXMrgX3P3kpOXD0JAr2/BryssfwNOrNc7pmLFrKvIH1lOrMEPjzpBVHdTXSeVW02aNIlJkyaZ9Zht/asy/skm7Dh5iQ+WHNJulrK1h/6zwasRLBwEifvMmkmpOKynyKefhTO7+f1acx5uVEPvNIqFCgsLIywszOzH7de8Dm88EMCCyNNM2Wq8+9XBDZ5bBM5Vta6VaoYpxQSsp8gfXgpoTTUPNfLSOYxiqTZs2MCGDRt0OfabD9bn8bBafLE2+p8+9G41tJul8nNhTl/ISNYlm2K9rKfIH1xIrG0AttUb4OfpqncaxUKNHTuWsWPH6nJsIQSf921CS58qjFq4n70JqdoKz/rwzALt0+i8pyE7Q5d8inWyjiKffBLO/k34tdbqLF6xaI52BqYObEFNd0denhXJqUvGu1/rtjbeLPU3LB6qbpZSyox1FPmDi5AIlue25aHGqj1esWweLvbMGNyS3HzJkJm7uZxp7FoZ3BMe/RKOr4VVo9TNUkqZKP9FXko4uIjD9k1w9fSmYc1KeidSlLvy83Rl2sDmnErJ5NU5e8nOzddWtHwROr4F+36FrV/oG1KxCuW/yJ/9G1JimJ3Rit6htdUEIUq50dqvKl/0a8LO2Ev/jEMP8MCH2s1SW8bBvln6hlTKvfI/nvzBReQJO9bktWJZmBqrW7mzqVOn6h2hWfBLsgAAEuFJREFUiD5N65BwKZNJG07gW82F17oGaDdL9f4Orl6AFSPB1QvqP6R3VKWcKt9n8vl5cOg3dtm1wKdOLXyrueidSLFwQUFBBAUF6R2jiBHdAunTtDZfrotmWVSittBgB0//CjUaw6LBcGavrhmV8qt8F/m4P+FqErOvtuLxsNp6p1HKgRUrVrBixQq9YxQhhOB/fUNo5evB24v2s/bQeW2Fgxs8uwhcPGHeU2oKQeWelO8if2Ah122c2Saa8YRqqlFKYOLEiUycOFHvGLdwsDXw06AWNK7tzuvz9rHm4DlthZsXDFyifT27j9aEoyilUH6L/LV05JGlrJLt6BjsTVVXB70TKcp9qeRox6yhrQj1rszr4X//c1dsVX94diFkXIS5/eD6FX2DKuVK+S3yh35D5GQy61onnm7hrXcaRSkTbo7aOPTN61bhjfC/mb/7lLaiTgt4aiacP6QNaJaXo2tOpfwov0X+79mctvPhvGsjOtVXwwor1sPVwZYZQ1rSMdCT934/yIR10Vr3yvoPQa9JELMRlv+fullKKZHyWeSTDkPiXmZkduTZNvUw2Ki+8Yp1cXGw5adBLRjQ0pvvN59k1ML92g1TzV6ALqNhfzhs/I/eMZVyoHz2k9/zE7nCnlV0ZGWrunqnUcqR2bNn6x2hxOwMNox/MgRvD2e+XBdNYmoWk59vRrXO78CVs7D9K6hUS5tSUFFuo/ydyWelIffPZ0V+O9o1CcLTTV1wVUrO29sbb+/ycw1HCMFrXQP4ZkAY+8+k0fu77RxMTIdHJ0LQo7D633Bkud4xFQtW/or8/nBETiY/ZT/IoHY+eqdRypkFCxawYMECvWOU2uNhtflteDuEEPT7MYKlB5K0USvrtIDfXoT47XpHVCxU+Sry+fnk757OfhGEm29zwrwr651IKWemTJnClClT9I5xTxrXdmfZ6+0J9a7MyAVR/Hd9PLn9w6GKD8zrD6f36B1RsUDlq8gfW4lNSgzTrj/E610D9U6jKGZXzdWBuS+25oW29Zi+LY7/b+/Mw6OqsgT+O6ktC4EEwqogO4jI3rQIo2mVVlGJn4g4rTNq9ziOtii96Nhu7dLt1q32uNHuiNKA4oYbamvjLhAWISwKiLLvAjEkBFKn/7g3pGSApKGSqlTO7/veV++euu+9806dOvXuufXOu2jycraM8HfFThgB6+YnWkUjyag/QV6V6Mf3s1pasbb1KQzu3CzRGhlGQggF0ritoCf3nNuLwm++47SnljI7fzyEs+HZs2HjkkSraCQR9SfIr/iAtLVzeLj8DEYP7W4lhY0Gz3kD2vLqlYPJTg8yctIqxnd9EE0LwbgzYP2CRKtnJAn1I8irUjH9HjaTw8q2BfykW4tEa2QYSUH3Vo157cohDO/dhps/LuXaRndQEYi4QG85eoP6EuSXv0dg5Sc8sLuA3wzrZVfxxiEzZcoUpkyZkmg14kpWJMj9o/pw94hjmbo6k+ElN1ISaIKOL4CvP0i0ekaCSf4gH42y6+3fs1qbs63HBfRrl5tojYx6TF5eHnl5eYlWI+6ICKN+1I7XRg8h2LQd+Vv+lzW0QCeMhC8mJVo9I4EkfZDXoilENhXxEKO4/szeiVbHqOeMGzeOcePGJVqNWqNry2xevPx4Lh02iHNKb2Dmns7w8mXsnnaje8iO0eBI7iBftp1db/yOedGOdB96Ca2apCdaI6Oek+pBHiAYSOO/T+jE5DHDeKrDvYzfM5TQ5w+ydmwBe0q+S7R6Rh2T1EG++M1bCJVtYULeGC4c1DHR6hhGvaJDXhaPXjyI7r94jLHZV9J846ds/VN/3nr5OXaUWanihkLSBvk9X39E1vynmcypXHXhSIKBpFXVMJKagR2a8j+//gNzh06mPJjF6V/8krfuPJ8/vjSDojXbXRljI2VJyiqUunMr30/8OVujLWlS8AfaNs1MtEqGUa8REQYOGQo/nsWmqTczcsFjbPhiDvcUjmJx3mmMGNCOgr5taJFtKdFUI/kuj6MVrHrqIjLLt/BRr7s4o7+VLzCMuBFKp/mIe0j7+ds0b92O+8NjebTkKhZOe5wT75zGxU/P5IXCVWzfaemcVEFqY6g2YMAALSwsPKRtFz99BUd/O4FJeVdx3hW3kWYPBDHiyM6dOwHIzLTRIdEoLHwJPvwzbFpMWSCbaQxi/M7jWZDWlcGdmzPs2Nac2qMVTTJDida2QSAis1V1QFz3mSxBXqNRCp+9nh+tGMu7jc/hxKueJBxMvoGGYaQc0SismA7zJqKLX0P2lFISzOUzPYb3y7oxW4+mdaeeDO3ZhhO6NLf0aS2SskG+tKycWY//khO2PM+M7KH0Gf03IuFw3PUyjEceeQSAK664IsGaJCm7imHx67D8ffSbj5DidQDsJJ2i6FEsjLZnfaPuNG4/gKO69aFfh+a0yclIsNKpQ0oG+XmLlhB98VL6Vcxnbqvz6H3pX0kLBOKuk2EA5OfnAzB9+vSE6lEvUIUty2DVTHTdPMpWziG4sYhQtAyAUg2zWNuxPNiF4tweRFv3IaddT7q0bkrnFo3IiiTl/zqSmtoI8gn7FBat3MCiV+/jp5ufISRRlg26m74/vQysLo1hJAcikNcF8rogfS8gA9xds5uXUrFmLiXLZ9FmzTx6bP+A9C1vwRbYtSDEYm3LK9EOrA61Z0d2B8pzupLV7AiOyM2kdU46rZtk0LxRhGaNwvZDUAfUmYVVlW83lzC/8EN2L3iJ/JJp9JBiVjQdTMuR99G5Tfe6UsUwjEMlLQAtuhNo0Z28vv/uZNEobP2aijVzKF0xm7Zr5tH9u5mk73kPdgA7YMe3mSzTNqzRPOZqLhv8si3QDMnMJZSVQ0ajXLKyc2nWOJ1mWe5HIDczTE5maO9ro0jQChT+i9RKkC8tLeXTD9+hdNsGyratZ8/m5TTesZSe+hXDZQcVpLGqxYlETr6aDt1/UhsqGIZRV6SlQV5nAnmdyel9npOpwvcbYNMS2PQV2ZuW0Gvjl/TcvobA918QqCit2n6XX7a6ZrFmUEwGxZpJGWHKCPGNhthFiN0SRgMRNBhBgumkhdMJhDJIC4VJC4YJBCMEQhEC4TChUIRgKJ1QOEIwHHHvhyJ7l2AoRDAUcUs4nWA4TDCUjgRCSDCMBMIpkVmolSCf8d0Sjn9/5N52BWlsibSlOC8f7ZZPi/4FtM9KvUqAhmF4RCC7lVs65iPEBBtV2LUDitdD8Too3ebaZTtg1w4alW0nXLKNxju3s6e8jIryUnR3GbqnFNmzDanYRaBiF4GyckKl5YTYTYBorZxGuQaIkoZSfbCvSZ9EUCsTryJSDHwZ9x3XT/KAzYlWIkkwW1RhtqjCbFFFN1XNjucOaysn/2W8Z4jrKyJSaLZwmC2qMFtUYbaoQkQO7S7Sg2B3GxmGYaQwFuQNwzBSmNoK8o/V0n7rI2aLKswWVZgtqjBbVBF3W9TKxKthGIaRHFi6xjAMI4WxIG8YhpHCHFaQF5HTRORLEVkmItd52QQRmS8id8T0u1FEzj5cZZMJEWkrIv8QkUUislBErvbypiLyrogs9a+5Xj7C9/tIRJp5WScRmZzI84gnIhIQkbki8rpvdxCRGd4/JotI2MtHi0iRiLwZIxsiIvcnUv94ISI5IjJFRJaIyGIRGdRQ/UJEfuXPr0hEJopIekPxCxF5SkQ2ikhRjOxAfiAi8oC3yXwR6efl3URktpcN8rKgiPxdRGpW81lVD2kBAsByoCMQBr4AegFP+PffBZoArYHXDvU4ybr48+rn17OBr4AewD3AdV5+HXC3X58OZAIXAqO9bCLQJdHnEkeb/Br4G/C6bz8PnO/X/wpc7tc/x11g3AicBQjwNtA00ecQJzs8A/yXXw8DOQ3RL4AjgBVARow/XNxQ/AI4AegHFMXIDuQHw4C3/DkfB8zw8vuAIcCRwIteNhq4uKZ6HM6V/EBgmap+rarlwCTgDCBDRNKAEFAB3Ab8/jCOk5So6jpVnePXi4HFOKcuwH3J8a+VI5goEMF9oXeLyL8B61V1aZ0qXkuIyJG4z/8J3xbgJGCK7xJrC8H5RyawGxfg3lLVrXWpc20gIk1wX+4nAVS1XFW30UD9AnfDZYaIBHHnuI4G4heq+iF7K/Ls5UB+UACMV8fnQI6ItMbZIZMq/8jB/QCOr6keh3PH6xHAqpj2auDHwCZgDvAs0BlIqwyGqYqItAf6AjOAlqq6zr+1Hmjp1+8E/g6sxTnvC8D5dapo7fIX4FrcqAagGbBNVff49mqczwA8hLtqWwh8ArwKnFp3qtYqHXDfgadFpDcwG7iaBugXqrpGRP4MrARKgXdw9miIflHJgfxgf/H0COBhXECPAJcBNwF3qGqNi/XEfeJVVceoah9VvRe4HbhJRG4QkedF5NJ4Hy/RiEgj4EVgjKruiH1P3dhK/fq7qtpfVc/C/Wq/CXT1udvHa5xfS0JE5Exgo6rOrkl/VX1WVfuq6oXAr4AHgNO9Le73I8H6ShA3RB+rqn2BEtywfC8NyC9ycefUAWgDZAGnHah/ivvF/yPWDw7SZ6Wq5qvqIGAnLm2zWESe9fMZXas7zuEYbQ3QNqZ9pJcBICIFuF/tRkAnVT0POLc+O+2+iEgIF+AnqOpLXrzBD7Pwrxv32SYTl5d8GLgVuAj4GLigjtSuDQYDw0XkG1za7iTg/3BDzsrR4g/8A0BE2gADVfUV4DfAKGAbcHId6V0brAZWq+oM356CC/oN0S9OAVao6iZV3Q28hPOVhugXlRzIDw4aTz1/xM1XXIVLi15LDVLhhxPkZwFd/Ex5GDfEnOqVDwFjcJMMGVT9WgVwE1H1Hp9zfhJYrKr3xbw1FfcFxb++us+m1wAPeKevtE0Ul3Orl6jq71T1SFVtj/OD91X1AuAfwLm+2/5scTtws19PFVusB1aJSDcvOhlYRAP0C1ya5jgRyfTfl0pbNDi/iOFAfjAV+E//L5vjgO0xaR1E5ERgrZ+rycTZo2Y2OczZ42G4f5UsB26IkY/Bz/7iJlMmAgvwM8mpsOBmvBWYD8zzyzBcLvo9YCku19o0Zps2wBsx7ZFU5R+bJ/qc4mSXfKr+XdMRmAksw+WaIzH9+gJP7uMzC4Fpsf3q4wL0AQq9b7wC5DZUv8CNSpYARbh5ukhD8Qsf99bhJk9XA784kB/4OPmwj6ULgAEx+xHcvxUr+x6Nm/ecDwyuTg8ra2AYhpHCpNREhmEYhvFDLMgbhmGkMBbkDcMwUhgL8oZhGCmMBXnDMIwUxoK8kXSIyDcikne4fWrr2Pv0v0VEfrsfeRsRmeLX86WqMudwqarYeraI9IiX7oaxPyzIGw2OmLstaw1VXauq5+5HPlVV7/LNs3GVSw2j1rAgbxw2ItJeXO30cSLylbhnCpwiIp/4utkDfb+mIvKKr439uYj08vJmIvKOuLrjT+Bu/qjc94UiMlNE5onIoyISqEaX732dk4Ui8p6INPfy6SLyFxEpBK4WkZPF1b5fIK7udyRmN9d6+UwR6ey3P0tcDfS5vpZ3y5j+vUXkM3+ul8bYpIh9EJGLReQhETkeGA78yZ9bJxGZE9OvS2zbMA4VC/JGvOgM3At098vPcHcF/xa43ve5FZirqr28rLJc6u+Bj1X1GOBloB2AiByNq1syWFX74EpXV1fLJQso9Pv6gB/W9gir6gDcnYXjgFGqeiyuqNjlMf22e/lDuOqa4OrIHKeu6NgkXN2QSnrh6vUMAm72tVcOiqp+iruV/Rp1Bf2WA9tFpI/vcgnwdHX7MYzqsCBvxIsVqrpAXQnUhcB76m6nXgC0932G4G5tR1XfB5qJSGNc/fXnvPwN4Dvf/2SgPzBLROb5dsdq9IgClU9Ves4fs5JKeTev71e+/YzXoZKJMa+D/PqRwNsisgBXZ+aYmP6vqmqpqm7G1WUZWI2OB+IJ4BI/WhmFewCLYRwWFuSNeLErZj0a045y6M8tEOAZf6XbR1W7qeot/+I+Yut2lBzCNpXrDwIP+Sv8y4D0A/TfX7umvAicDpwJzFbVLYe4H8PYiwV5oy75CJ9uEZF8YLO6Gvwf4tI7iMjpuIJe4Ao5nSsiLfx7TUXkqGqOkUZVhcOf4dIs+/Il0L4y3w78By61U8momNfP/HoTqkq/XsQPKRD37NJmuAJts6rRsZJiqh6ygqqW4R53NxZL1RhxwoK8UZfcAvQXkfnAXVQFy1uBE0RkIXAOrkQtqroIVz/7Hb/Nu7hn6x6MEmCgn/Q8Cff4yR/gg+klwAs+/RLFPWu0klx/vKtxD6+o1P0FEZkNbN5nl/NxaZrPgdtVdW01OlYyCbjGT+Z28rIJXp93argPwzgoVoXSSClE5HtVbZRoPQ4V/5/7Jqp6U6J1MVKDWv+/sGEYNUNEXgY64UYghhEX7EreMAwjhbGcvGEYRgpjQd4wDCOFsSBvGIaRwliQNwzDSGEsyBuGYaQw/wTyKB5ceZDkZwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the x only distribution\n",
    "dist_plot_save(kde_df, value_col = 'dist_x',\n",
    "               plt_title = 'Logistic regression model for y=x',\n",
    "               out_fn = os.path.join(local_save_path, 'PLOT_model_prob_dist_x.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Plot the \"actual\" probability by race/ethinicity (input simulated probability)\n",
    "This can be compared to the prob_x_age distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Actual KDE alone\n",
    "kde_actual_df = pd.concat([kde_est_race_eth(score_dist_x, 'WHITE ALONE, NOT HISPANIC OR LATINO', \n",
    "                                            pop_data, 'sim_prob'),\n",
    "                          kde_est_race_eth(score_dist_x, 'HISPANIC OR LATINO', \n",
    "                                           pop_data, 'sim_prob')], axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "dist_plot_save(kde_actual_df, value_col = 'dist',\n",
    "               plt_title = 'Input (\"actual\") probability',\n",
    "               x_label = 'target probability')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fairness Metrics\n",
    "I focus on just 3 metrics: False positives, false negatives, and accuracy.  I look at raw metrics, as well as metrics by age category"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Functions to calculate metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "      <th>x</th>\n",
       "      <th>eta</th>\n",
       "      <th>sim_lin</th>\n",
       "      <th>sim_prob</th>\n",
       "      <th>targ_status</th>\n",
       "      <th>prob_x_age</th>\n",
       "      <th>prob_x</th>\n",
       "      <th>targ_x_age</th>\n",
       "      <th>targ_x</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>1.624345</td>\n",
       "      <td>-0.122474</td>\n",
       "      <td>1.112098</td>\n",
       "      <td>0.752520</td>\n",
       "      <td>1</td>\n",
       "      <td>0.775692</td>\n",
       "      <td>0.571468</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>-0.611756</td>\n",
       "      <td>0.228170</td>\n",
       "      <td>-1.088939</td>\n",
       "      <td>0.251818</td>\n",
       "      <td>0</td>\n",
       "      <td>0.242198</td>\n",
       "      <td>0.237208</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>52.0</td>\n",
       "      <td>-0.528172</td>\n",
       "      <td>-0.352305</td>\n",
       "      <td>-1.363402</td>\n",
       "      <td>0.203688</td>\n",
       "      <td>0</td>\n",
       "      <td>0.205344</td>\n",
       "      <td>0.247195</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>20.0</td>\n",
       "      <td>-1.072969</td>\n",
       "      <td>-0.830553</td>\n",
       "      <td>-5.156024</td>\n",
       "      <td>0.005732</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005765</td>\n",
       "      <td>0.187196</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>69.0</td>\n",
       "      <td>0.865408</td>\n",
       "      <td>-0.261090</td>\n",
       "      <td>1.739299</td>\n",
       "      <td>0.850598</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862929</td>\n",
       "      <td>0.448610</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth  age_yr         x       eta   sim_lin  \\\n",
       "0  WHITE ALONE, NOT HISPANIC OR LATINO    55.0  1.624345 -0.122474  1.112098   \n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO    55.0 -0.611756  0.228170 -1.088939   \n",
       "2  WHITE ALONE, NOT HISPANIC OR LATINO    52.0 -0.528172 -0.352305 -1.363402   \n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO    20.0 -1.072969 -0.830553 -5.156024   \n",
       "4  WHITE ALONE, NOT HISPANIC OR LATINO    69.0  0.865408 -0.261090  1.739299   \n",
       "\n",
       "   sim_prob  targ_status  prob_x_age    prob_x  targ_x_age  targ_x  \n",
       "0  0.752520            1    0.775692  0.571468           1       1  \n",
       "1  0.251818            0    0.242198  0.237208           0       0  \n",
       "2  0.203688            0    0.205344  0.247195           0       0  \n",
       "3  0.005732            0    0.005765  0.187196           0       0  \n",
       "4  0.850598            1    0.862929  0.448610           1       0  "
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_pred_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "def metric_fp(data, model_target_col = 'targ_x_age'):\n",
    "    \"\"\" Get false positive rate.  Use the model prediction in the indicated column.\n",
    "    Inputs:\n",
    "        data: Data for the population of interest, with predictions (model inputs and outputs)\n",
    "        model_target_col: Name of the model prediction column\"\"\"\n",
    "    sel_d = np.where(data['targ_status'] == 0, True, False)\n",
    "    fp_ser = np.where(data[sel_d][model_target_col] == 1, 1, 0)\n",
    "    try:\n",
    "        fp_rate = np.mean(fp_ser)\n",
    "    except:\n",
    "        fp_rate = None\n",
    "    return pd.DataFrame.from_dict({'fp_rate': [fp_rate]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "def metric_fn(data, model_target_col = 'targ_x_age'):\n",
    "    \"\"\" Get false negative rate.  Use the model prediction in the indicated column.\n",
    "    Inputs:\n",
    "        data: Data for the population of interest, with predictions (model inputs and outputs)\n",
    "        model_target_col: Name of the model prediction column\"\"\"\n",
    "    sel_d = np.where(data['targ_status'] == 1, True, False)\n",
    "    fn_ser = np.where(data[sel_d][model_target_col] == 0, 1, 0)\n",
    "    try:\n",
    "        fn_rate = np.mean(fn_ser)\n",
    "    except:\n",
    "        fn_rate = None\n",
    "    return pd.DataFrame.from_dict({'fn_rate': [fn_rate]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "def metric_accuracy(data, model_target_col = 'targ_x_age'):\n",
    "    \"\"\" Get accuracy.  Use the model prediction in the indicated column.\n",
    "    Inputs:\n",
    "        data: Data for the population of interest, with predictions (model inputs and outputs)\n",
    "        model_target_col: Name of the model prediction column\"\"\"\n",
    "    acc_ser = np.where(data['targ_status'] == data[model_target_col], 1, 0)\n",
    "    try:\n",
    "        accuracy = np.mean(acc_ser)\n",
    "    except:\n",
    "        accuracy = None\n",
    "    return pd.DataFrame.from_dict({'accuracy': [accuracy]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fp_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.105271</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    fp_rate\n",
       "0  0.105271"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metric_fp(pop_pred_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fp_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>race_eth</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HISPANIC OR LATINO</th>\n",
       "      <td>0.069099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WHITE ALONE, NOT HISPANIC OR LATINO</th>\n",
       "      <td>0.148039</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      fp_rate\n",
       "race_eth                                     \n",
       "HISPANIC OR LATINO                   0.069099\n",
       "WHITE ALONE, NOT HISPANIC OR LATINO  0.148039"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_pred_data.groupby('race_eth').apply(lambda x: metric_fp(x)).reset_index(level=1, drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics by race/ethnicity group only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_x_age = pd.concat([pop_pred_data.groupby('race_eth').apply(lambda x: metric_fp(x)),\n",
    "                          pop_pred_data.groupby('race_eth').apply(lambda x: metric_fn(x)),\n",
    "                          pop_pred_data.groupby('race_eth').apply(lambda x: metric_accuracy(x))], axis=1) \\\n",
    "    .reset_index(level=1, drop=True)\n",
    "metrics_x_age['model'] = 'x_age'\n",
    "metrics_x = pd.concat([pop_pred_data.groupby('race_eth').apply(lambda x: metric_fp(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby('race_eth').apply(lambda x: metric_fn(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby('race_eth').apply(lambda x: metric_accuracy(x, 'targ_x'))], axis=1) \\\n",
    "    .reset_index(level=1, drop=True)\n",
    "metrics_x['model'] = 'x'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>fp_rate</th>\n",
       "      <th>fn_rate</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.069099</td>\n",
       "      <td>0.407381</td>\n",
       "      <td>0.8374</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.148039</td>\n",
       "      <td>0.257732</td>\n",
       "      <td>0.8094</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.075732</td>\n",
       "      <td>0.775687</td>\n",
       "      <td>0.7308</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.075817</td>\n",
       "      <td>0.809278</td>\n",
       "      <td>0.6396</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth   fp_rate   fn_rate  accuracy  model\n",
       "0                   HISPANIC OR LATINO  0.069099  0.407381    0.8374  x_age\n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO  0.148039  0.257732    0.8094  x_age\n",
       "2                   HISPANIC OR LATINO  0.075732  0.775687    0.7308      x\n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO  0.075817  0.809278    0.6396      x"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics_race_eth = pd.concat([metrics_x_age, metrics_x]).reset_index()\n",
    "metrics_race_eth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_race_eth.to_csv(os.path.join(local_save_path, 'DATA_metrics_by_race_eth.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics by race/ethnicity plus age category\n",
    "See if metrics \"pass\" when we add an age group.  Use large age buckets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_cond = [pop_pred_data.age_yr < 30,\n",
    "           (pop_pred_data.age_yr >= 30) & (pop_pred_data.age_yr < 50),\n",
    "           (pop_pred_data.age_yr >= 50) & (pop_pred_data.age_yr < 70),\n",
    "            (pop_pred_data.age_yr >= 70) & (pop_pred_data.age_yr < 80),\n",
    "           pop_pred_data.age_yr >= 80]\n",
    "age_cond_lbl = ['18-29', '30-49', '50-69', '70-79', '80+']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_cond = [pop_pred_data.age_yr < 40,\n",
    "           (pop_pred_data.age_yr >= 40) & (pop_pred_data.age_yr < 70),\n",
    "           pop_pred_data.age_yr >= 70]\n",
    "age_cond_lbl = ['18-39', '40-69','70+']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "pop_pred_data['age_category'] = np.select(age_cond, age_cond_lbl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_cat_x_age = pd.concat([pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_fp(x)),\n",
    "                          pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_fn(x)),\n",
    "                          pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_accuracy(x))], axis=1) \\\n",
    "    .reset_index(level=2, drop=True)\n",
    "metrics_cat_x_age['model'] = 'x_age'\n",
    "metrics_cat_x = pd.concat([pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_fp(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_fn(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby(['age_category', 'race_eth']).apply(lambda x: metric_accuracy(x, 'targ_x'))], axis=1) \\\n",
    "    .reset_index(level=2, drop=True)\n",
    "metrics_cat_x['model'] = 'x'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age_category</th>\n",
       "      <th>race_eth</th>\n",
       "      <th>fp_rate</th>\n",
       "      <th>fn_rate</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>18-39</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.002352</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.937004</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18-39</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.002053</td>\n",
       "      <td>0.952381</td>\n",
       "      <td>0.934227</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40-69</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.128278</td>\n",
       "      <td>0.471186</td>\n",
       "      <td>0.739547</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40-69</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.202867</td>\n",
       "      <td>0.382716</td>\n",
       "      <td>0.723278</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>70+</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.790123</td>\n",
       "      <td>0.022663</td>\n",
       "      <td>0.834101</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>70+</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.818627</td>\n",
       "      <td>0.032445</td>\n",
       "      <td>0.817245</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>18-39</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.089370</td>\n",
       "      <td>0.520833</td>\n",
       "      <td>0.883260</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>18-39</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.102669</td>\n",
       "      <td>0.685714</td>\n",
       "      <td>0.858238</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>40-69</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.058115</td>\n",
       "      <td>0.772881</td>\n",
       "      <td>0.666376</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>40-69</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.055197</td>\n",
       "      <td>0.799383</td>\n",
       "      <td>0.639206</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>70+</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.024691</td>\n",
       "      <td>0.886686</td>\n",
       "      <td>0.274194</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>70+</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.024510</td>\n",
       "      <td>0.835458</td>\n",
       "      <td>0.319588</td>\n",
       "      <td>x</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age_category                             race_eth   fp_rate   fn_rate  \\\n",
       "0         18-39                   HISPANIC OR LATINO  0.002352  0.958333   \n",
       "1         18-39  WHITE ALONE, NOT HISPANIC OR LATINO  0.002053  0.952381   \n",
       "2         40-69                   HISPANIC OR LATINO  0.128278  0.471186   \n",
       "3         40-69  WHITE ALONE, NOT HISPANIC OR LATINO  0.202867  0.382716   \n",
       "4           70+                   HISPANIC OR LATINO  0.790123  0.022663   \n",
       "5           70+  WHITE ALONE, NOT HISPANIC OR LATINO  0.818627  0.032445   \n",
       "6         18-39                   HISPANIC OR LATINO  0.089370  0.520833   \n",
       "7         18-39  WHITE ALONE, NOT HISPANIC OR LATINO  0.102669  0.685714   \n",
       "8         40-69                   HISPANIC OR LATINO  0.058115  0.772881   \n",
       "9         40-69  WHITE ALONE, NOT HISPANIC OR LATINO  0.055197  0.799383   \n",
       "10          70+                   HISPANIC OR LATINO  0.024691  0.886686   \n",
       "11          70+  WHITE ALONE, NOT HISPANIC OR LATINO  0.024510  0.835458   \n",
       "\n",
       "    accuracy  model  \n",
       "0   0.937004  x_age  \n",
       "1   0.934227  x_age  \n",
       "2   0.739547  x_age  \n",
       "3   0.723278  x_age  \n",
       "4   0.834101  x_age  \n",
       "5   0.817245  x_age  \n",
       "6   0.883260      x  \n",
       "7   0.858238      x  \n",
       "8   0.666376      x  \n",
       "9   0.639206      x  \n",
       "10  0.274194      x  \n",
       "11  0.319588      x  "
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics_race_eth_cat = pd.concat([metrics_cat_x_age, metrics_cat_x]).reset_index()\n",
    "metrics_race_eth_cat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_race_eth_cat.to_csv(os.path.join(local_save_path, 'DATA_metrics_by_race_eth_age_category.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics by finer-grained age/race/ethnicity\n",
    "Get within-5-year buckets, atarting at 18.  Set group to midpoint of bucket."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "      <th>x</th>\n",
       "      <th>eta</th>\n",
       "      <th>sim_lin</th>\n",
       "      <th>sim_prob</th>\n",
       "      <th>targ_status</th>\n",
       "      <th>prob_x_age</th>\n",
       "      <th>prob_x</th>\n",
       "      <th>targ_x_age</th>\n",
       "      <th>targ_x</th>\n",
       "      <th>age_category</th>\n",
       "      <th>age_grp_5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9995</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.072959</td>\n",
       "      <td>-0.943208</td>\n",
       "      <td>-3.021362</td>\n",
       "      <td>0.046470</td>\n",
       "      <td>0</td>\n",
       "      <td>0.050952</td>\n",
       "      <td>0.326900</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>18-39</td>\n",
       "      <td>30.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9996</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>50.0</td>\n",
       "      <td>0.489102</td>\n",
       "      <td>0.782216</td>\n",
       "      <td>-0.432676</td>\n",
       "      <td>0.393487</td>\n",
       "      <td>1</td>\n",
       "      <td>0.384404</td>\n",
       "      <td>0.389053</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>50.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9997</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>61.0</td>\n",
       "      <td>-1.014144</td>\n",
       "      <td>2.408434</td>\n",
       "      <td>-0.673300</td>\n",
       "      <td>0.337758</td>\n",
       "      <td>1</td>\n",
       "      <td>0.275669</td>\n",
       "      <td>0.193093</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>60.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9998</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>47.0</td>\n",
       "      <td>-0.062696</td>\n",
       "      <td>0.882786</td>\n",
       "      <td>-1.274418</td>\n",
       "      <td>0.218502</td>\n",
       "      <td>1</td>\n",
       "      <td>0.204220</td>\n",
       "      <td>0.307770</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>45.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9999</th>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>60.0</td>\n",
       "      <td>-1.437870</td>\n",
       "      <td>-0.099596</td>\n",
       "      <td>-1.447830</td>\n",
       "      <td>0.190336</td>\n",
       "      <td>0</td>\n",
       "      <td>0.179781</td>\n",
       "      <td>0.153694</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>60.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                race_eth  age_yr         x       eta   sim_lin  sim_prob  \\\n",
       "9995  HISPANIC OR LATINO    30.0  0.072959 -0.943208 -3.021362  0.046470   \n",
       "9996  HISPANIC OR LATINO    50.0  0.489102  0.782216 -0.432676  0.393487   \n",
       "9997  HISPANIC OR LATINO    61.0 -1.014144  2.408434 -0.673300  0.337758   \n",
       "9998  HISPANIC OR LATINO    47.0 -0.062696  0.882786 -1.274418  0.218502   \n",
       "9999  HISPANIC OR LATINO    60.0 -1.437870 -0.099596 -1.447830  0.190336   \n",
       "\n",
       "      targ_status  prob_x_age    prob_x  targ_x_age  targ_x age_category  \\\n",
       "9995            0    0.050952  0.326900           0       0        18-39   \n",
       "9996            1    0.384404  0.389053           0       0        40-69   \n",
       "9997            1    0.275669  0.193093           0       0        40-69   \n",
       "9998            1    0.204220  0.307770           0       0        40-69   \n",
       "9999            0    0.179781  0.153694           0       0        40-69   \n",
       "\n",
       "      age_grp_5  \n",
       "9995       30.5  \n",
       "9996       50.5  \n",
       "9997       60.5  \n",
       "9998       45.5  \n",
       "9999       60.5  "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_pred_data['age_grp_5'] = 18 + 5 * np.floor((pop_pred_data['age_yr'] - 18)/5) + 2.5\n",
    "pop_pred_data.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda/envs/azureml_py36/lib/python3.6/site-packages/numpy/core/fromnumeric.py:3335: RuntimeWarning: Mean of empty slice.\n",
      "  out=out, **kwargs)\n",
      "/anaconda/envs/azureml_py36/lib/python3.6/site-packages/numpy/core/_methods.py:161: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  ret = ret.dtype.type(ret / rcount)\n"
     ]
    }
   ],
   "source": [
    "metrics_grp5_x_age = pd.concat([pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_fp(x)),\n",
    "                          pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_fn(x)),\n",
    "                          pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_accuracy(x))], axis=1) \\\n",
    "    .reset_index(level=2, drop=True)\n",
    "metrics_grp5_x_age['model'] = 'x_age'\n",
    "metrics_grp5_x = pd.concat([pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_fp(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_fn(x, 'targ_x')),\n",
    "                          pop_pred_data.groupby(['age_grp_5', 'race_eth']).apply(lambda x: metric_accuracy(x, 'targ_x'))], axis=1) \\\n",
    "    .reset_index(level=2, drop=True)\n",
    "metrics_grp5_x['model'] = 'x'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age_grp_5</th>\n",
       "      <th>race_eth</th>\n",
       "      <th>fp_rate</th>\n",
       "      <th>fn_rate</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20.5</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.961749</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20.5</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.975000</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>25.5</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.003817</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.957798</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>25.5</td>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.947802</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>30.5</td>\n",
       "      <td>HISPANIC OR LATINO</td>\n",
       "      <td>0.001992</td>\n",
       "      <td>0.931034</td>\n",
       "      <td>0.947269</td>\n",
       "      <td>x_age</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age_grp_5                             race_eth   fp_rate   fn_rate  \\\n",
       "0       20.5                   HISPANIC OR LATINO  0.000000  1.000000   \n",
       "1       20.5  WHITE ALONE, NOT HISPANIC OR LATINO  0.000000  1.000000   \n",
       "2       25.5                   HISPANIC OR LATINO  0.003817  1.000000   \n",
       "3       25.5  WHITE ALONE, NOT HISPANIC OR LATINO  0.000000  1.000000   \n",
       "4       30.5                   HISPANIC OR LATINO  0.001992  0.931034   \n",
       "\n",
       "   accuracy  model  \n",
       "0  0.961749  x_age  \n",
       "1  0.975000  x_age  \n",
       "2  0.957798  x_age  \n",
       "3  0.947802  x_age  \n",
       "4  0.947269  x_age  "
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics_race_eth_grp5 = pd.concat([metrics_grp5_x_age, metrics_grp5_x]).reset_index()\n",
    "metrics_race_eth_grp5.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Plot metrics for the fine-grained age groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to plot and save \n",
    "def metric_plot_save(data, model_sel = 'x_age',\n",
    "                     index_col = 'age_grp_5',\n",
    "                     value_col = 'accuracy',\n",
    "                     metric_name = None,\n",
    "                     plt_title = 'Model y = x + age',\n",
    "                     out_fn = None):\n",
    "    \"\"\" Plots the metric vs. age group midpoint for one model \"\"\"\n",
    "    \n",
    "    mdl_data = data[data['model'] == model_sel]\n",
    "    maxv = np.max(mdl_data[value_col]) * 1.05\n",
    "    minv = np.min(mdl_data[value_col]) * 0.9\n",
    "    pivot_data = mdl_data.pivot(index=index_col, columns='race_eth', values=value_col)\n",
    "    labels = [ '\\n'.join(wrap(l, 20)) for l in pivot_data.columns]\n",
    "    \n",
    "    # Use metric name for title if unknown\n",
    "    if (metric_name is not None):\n",
    "        plt_suptitle = metric_name + ' vs. age group'\n",
    "    else:\n",
    "        plt_suptitle = value_col.capitalize() + ' vs. age group'\n",
    "    \n",
    "    plt.figure()\n",
    "    pivot_data.plot()\n",
    "    plt.gca().set_ylim(minv, maxv)\n",
    "    plt.legend(title=None, framealpha=1, labels=labels)\n",
    "    plt.xlabel('age group midpoint')\n",
    "    #plt.tick_params(axis='y', which='both', left=False, labelleft=False)\n",
    "    plt.gca().yaxis.set_major_formatter(matplotlib.ticker.PercentFormatter(1))\n",
    "    plt.ylabel(value_col)\n",
    "    plt.suptitle(plt_suptitle)\n",
    "    plt.title(plt_title)\n",
    "    if (out_fn is not None):\n",
    "        plt.savefig(out_fn, bbox_inches='tight')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "# Model y=x+age\n",
    "metric_plot_save(metrics_race_eth_grp5, out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_age_accuracy.png'))\n",
    "metric_plot_save(metrics_race_eth_grp5, value_col='fp_rate', metric_name = 'False positive rate',\n",
    "                 out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_age_fp_rate.png'))\n",
    "metric_plot_save(metrics_race_eth_grp5, value_col='fn_rate', metric_name = 'False negative rate',\n",
    "                 out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_age_fn_rate.png'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "# Model y=x\n",
    "metric_plot_save(metrics_race_eth_grp5, model_sel = 'x', plt_title='Model y = x',\n",
    "                 out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_accuracy.png'))\n",
    "metric_plot_save(metrics_race_eth_grp5, model_sel = 'x', plt_title='Model y = x',\n",
    "                 value_col='fp_rate', metric_name = 'False positive rate',\n",
    "                 out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_fp_rate.png'))\n",
    "metric_plot_save(metrics_race_eth_grp5, model_sel = 'x', plt_title='Model y = x',\n",
    "                 value_col='fn_rate', metric_name = 'False negative rate',\n",
    "                 out_fn=os.path.join(local_save_path, 'PLOT_metric_age_grp_x_fn_rate.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calibration\n",
    "Show calibration curve across groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>race_eth</th>\n",
       "      <th>age_yr</th>\n",
       "      <th>x</th>\n",
       "      <th>eta</th>\n",
       "      <th>sim_lin</th>\n",
       "      <th>sim_prob</th>\n",
       "      <th>targ_status</th>\n",
       "      <th>prob_x_age</th>\n",
       "      <th>prob_x</th>\n",
       "      <th>targ_x_age</th>\n",
       "      <th>targ_x</th>\n",
       "      <th>age_category</th>\n",
       "      <th>age_grp_5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>1.624345</td>\n",
       "      <td>-0.122474</td>\n",
       "      <td>1.112098</td>\n",
       "      <td>0.752520</td>\n",
       "      <td>1</td>\n",
       "      <td>0.775692</td>\n",
       "      <td>0.571468</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>40-69</td>\n",
       "      <td>55.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>55.0</td>\n",
       "      <td>-0.611756</td>\n",
       "      <td>0.228170</td>\n",
       "      <td>-1.088939</td>\n",
       "      <td>0.251818</td>\n",
       "      <td>0</td>\n",
       "      <td>0.242198</td>\n",
       "      <td>0.237208</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>55.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>52.0</td>\n",
       "      <td>-0.528172</td>\n",
       "      <td>-0.352305</td>\n",
       "      <td>-1.363402</td>\n",
       "      <td>0.203688</td>\n",
       "      <td>0</td>\n",
       "      <td>0.205344</td>\n",
       "      <td>0.247195</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>50.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>20.0</td>\n",
       "      <td>-1.072969</td>\n",
       "      <td>-0.830553</td>\n",
       "      <td>-5.156024</td>\n",
       "      <td>0.005732</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005765</td>\n",
       "      <td>0.187196</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>18-39</td>\n",
       "      <td>20.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WHITE ALONE, NOT HISPANIC OR LATINO</td>\n",
       "      <td>69.0</td>\n",
       "      <td>0.865408</td>\n",
       "      <td>-0.261090</td>\n",
       "      <td>1.739299</td>\n",
       "      <td>0.850598</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862929</td>\n",
       "      <td>0.448610</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>40-69</td>\n",
       "      <td>70.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              race_eth  age_yr         x       eta   sim_lin  \\\n",
       "0  WHITE ALONE, NOT HISPANIC OR LATINO    55.0  1.624345 -0.122474  1.112098   \n",
       "1  WHITE ALONE, NOT HISPANIC OR LATINO    55.0 -0.611756  0.228170 -1.088939   \n",
       "2  WHITE ALONE, NOT HISPANIC OR LATINO    52.0 -0.528172 -0.352305 -1.363402   \n",
       "3  WHITE ALONE, NOT HISPANIC OR LATINO    20.0 -1.072969 -0.830553 -5.156024   \n",
       "4  WHITE ALONE, NOT HISPANIC OR LATINO    69.0  0.865408 -0.261090  1.739299   \n",
       "\n",
       "   sim_prob  targ_status  prob_x_age    prob_x  targ_x_age  targ_x  \\\n",
       "0  0.752520            1    0.775692  0.571468           1       1   \n",
       "1  0.251818            0    0.242198  0.237208           0       0   \n",
       "2  0.203688            0    0.205344  0.247195           0       0   \n",
       "3  0.005732            0    0.005765  0.187196           0       0   \n",
       "4  0.850598            1    0.862929  0.448610           1       0   \n",
       "\n",
       "  age_category  age_grp_5  \n",
       "0        40-69       55.5  \n",
       "1        40-69       55.5  \n",
       "2        40-69       50.5  \n",
       "3        18-39       20.5  \n",
       "4        40-69       70.5  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pop_pred_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to create 2 calibration plots\n",
    "def calibration_decile_plot(data, \n",
    "                              model_prob = 'prob_x_age',\n",
    "                              plt_title = 'Model y = x + age',\n",
    "                              out_fn = None):\n",
    "    \"\"\" Calibration by race/ethnicity \"\"\"\n",
    "    \n",
    "    data['model_decile'] = pd.qcut(data[model_prob], q=10, labels=[str(x) for x in range(0,10)])\n",
    "    data_agg = data.groupby(['model_decile', 'race_eth']).agg(model_prob = (model_prob, 'mean'),\n",
    "                                                                actual_prob = ('targ_status', 'mean')) \\\n",
    "        .reset_index() \n",
    "    data_agg_pivot = data_agg.pivot(index='model_decile', columns='race_eth', values=['model_prob']) \\\n",
    "        .droplevel(axis=1, level=0)\n",
    "    labels = [ '\\n'.join(wrap(l, 20)) for l in data_agg_pivot.columns]\n",
    "    plt.figure()\n",
    "    data_agg_pivot.plot(style=['-','--'])\n",
    "    plt.legend(title=None, framealpha=1, labels=labels)\n",
    "    plt.xlabel('Model probability decile')\n",
    "    plt.gca().yaxis.set_major_formatter(matplotlib.ticker.PercentFormatter(1))\n",
    "    plt.ylabel('Actual rate')\n",
    "    plt.title(plt_title)\n",
    "    if (out_fn is not None):\n",
    "        plt.savefig(out_fn, bbox_inches='tight')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "calibration_decile_plot(pop_pred_data, out_fn = os.path.join(local_save_path, 'PLOT_calibration_deciles_age_x.png'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "calibration_decile_plot(pop_pred_data, 'prob_x',\n",
    "                         'Model y = x', out_fn = os.path.join(local_save_path, 'PLOT_calibration_deciles_x.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model probability vs. Actual probability (overall)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calibration_prob_align_plot(data, \n",
    "                              model_prob = 'prob_x_age',\n",
    "                              plt_title = 'Model y = x + age',\n",
    "                              out_fn = None):\n",
    "    \"\"\" Model probability vs. actual probability by race/ethnicity.  Uses\n",
    "    deciles to calculate model probabilities\"\"\"\n",
    "    \n",
    "    data['model_grp_decile'] = data.groupby('race_eth')[model_prob].transform(lambda x: pd.qcut(x, q=10, labels=[str(x) for x in range(0,10)]))\n",
    "\n",
    "    data_agg = data.groupby(['model_grp_decile', 'race_eth']).agg(model_prob = (model_prob, 'mean'),\n",
    "                                                                actual_prob = ('targ_status', 'mean')) \\\n",
    "        .reset_index() \n",
    "    labels = [ '\\n'.join(wrap(l, 20)) for l in data_agg['race_eth'].unique()]\n",
    "    \n",
    "    plt.figure()\n",
    "    fig, ax = plt.subplots()\n",
    "    data_agg.groupby('race_eth').plot(x='model_prob', y='actual_prob', ax=ax)\n",
    "    plt.legend(title=None, framealpha=1, labels=labels)\n",
    "    plt.xlabel('Model probability')\n",
    "    plt.gca().yaxis.set_major_formatter(matplotlib.ticker.PercentFormatter(1))\n",
    "    plt.gca().xaxis.set_major_formatter(matplotlib.ticker.PercentFormatter(1))\n",
    "    plt.ylabel('Actual rate')\n",
    "    plt.title(plt_title)\n",
    "    if (out_fn is not None):\n",
    "        plt.savefig(out_fn, bbox_inches='tight')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "calibration_prob_align_plot(pop_pred_data, out_fn = os.path.join(local_save_path, 'PLOT_calibration_probs_age_x.png'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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": [
    "calibration_prob_align_plot(pop_pred_data, 'prob_x',\n",
    "                            'Model y = x', out_fn= os.path.join(local_save_path, 'PLOT_calibration_probs_x.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusions\n",
    "- For processes that depend on age but not race/ethnicity, fairness metrics differ across racial/ethnic groups if age distributions vary.\n",
    "- Differences disappear when we examine fine-grained age categories.\n",
    "- Larger age buckets (10+ years) reduce but do not eliminate differences "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.6 - AzureML",
   "language": "python",
   "name": "python3-azureml"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}