notebooks/chexpert.sample.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Study population"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>patient_id</th>\n",
       "      <th>race</th>\n",
       "      <th>ethnicity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>patient48289</td>\n",
       "      <td>Other</td>\n",
       "      <td>Hispanic/Latino</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>patient33856</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>patient41673</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>patient48493</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     patient_id     race                ethnicity\n",
       "0  patient24428    White  Non-Hispanic/Non-Latino\n",
       "1  patient48289    Other          Hispanic/Latino\n",
       "2  patient33856    White  Non-Hispanic/Non-Latino\n",
       "3  patient41673  Unknown                  Unknown\n",
       "4  patient48493    White  Non-Hispanic/Non-Latino"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dir = '../datafiles/chexpert/'\n",
    "df_demo = pd.DataFrame(pd.read_excel(data_dir + 'CHEXPERT DEMO.xlsx', engine='openpyxl'))\n",
    "\n",
    "df_demo = df_demo.rename(columns={'PRIMARY_RACE': 'race'})\n",
    "df_demo = df_demo.rename(columns={'PATIENT': 'patient_id'})\n",
    "df_demo = df_demo.rename(columns={'GENDER': 'sex'})\n",
    "df_demo = df_demo.rename(columns={'AGE_AT_CXR': 'age'})\n",
    "df_demo = df_demo.rename(columns={'ETHNICITY': 'ethnicity'})\n",
    "df_demo = df_demo.drop(['sex', 'age'], axis=1)\n",
    "df_demo.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Path</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>Frontal/Lateral</th>\n",
       "      <th>AP/PA</th>\n",
       "      <th>No Finding</th>\n",
       "      <th>Enlarged Cardiomediastinum</th>\n",
       "      <th>Cardiomegaly</th>\n",
       "      <th>Lung Opacity</th>\n",
       "      <th>Lung Lesion</th>\n",
       "      <th>...</th>\n",
       "      <th>Pneumonia</th>\n",
       "      <th>Atelectasis</th>\n",
       "      <th>Pneumothorax</th>\n",
       "      <th>Pleural Effusion</th>\n",
       "      <th>Pleural Other</th>\n",
       "      <th>Fracture</th>\n",
       "      <th>Support Devices</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>split</th>\n",
       "      <th>patient_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CheXpert-v1.0/train/patient00002/study2/view1_...</td>\n",
       "      <td>Female</td>\n",
       "      <td>87</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>train</td>\n",
       "      <td>patient00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CheXpert-v1.0/train/patient00002/study1/view1_...</td>\n",
       "      <td>Female</td>\n",
       "      <td>83</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>train</td>\n",
       "      <td>patient00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CheXpert-v1.0/train/patient00003/study1/view1_...</td>\n",
       "      <td>Male</td>\n",
       "      <td>41</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>train</td>\n",
       "      <td>patient00003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>CheXpert-v1.0/train/patient00004/study1/view1_...</td>\n",
       "      <td>Female</td>\n",
       "      <td>20</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>PA</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>train</td>\n",
       "      <td>patient00004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>CheXpert-v1.0/train/patient00005/study1/view1_...</td>\n",
       "      <td>Male</td>\n",
       "      <td>33</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>PA</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>train</td>\n",
       "      <td>patient00005</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                Path     sex  age  \\\n",
       "1  CheXpert-v1.0/train/patient00002/study2/view1_...  Female   87   \n",
       "2  CheXpert-v1.0/train/patient00002/study1/view1_...  Female   83   \n",
       "4  CheXpert-v1.0/train/patient00003/study1/view1_...    Male   41   \n",
       "5  CheXpert-v1.0/train/patient00004/study1/view1_...  Female   20   \n",
       "7  CheXpert-v1.0/train/patient00005/study1/view1_...    Male   33   \n",
       "\n",
       "  Frontal/Lateral AP/PA  No Finding  Enlarged Cardiomediastinum  Cardiomegaly  \\\n",
       "1         Frontal    AP         NaN                         NaN          -1.0   \n",
       "2         Frontal    AP         NaN                         NaN           NaN   \n",
       "4         Frontal    AP         NaN                         NaN           NaN   \n",
       "5         Frontal    PA         1.0                         0.0           NaN   \n",
       "7         Frontal    PA         1.0                         NaN           0.0   \n",
       "\n",
       "   Lung Opacity  Lung Lesion  ...  Pneumonia  Atelectasis  Pneumothorax  \\\n",
       "1           1.0          NaN  ...        NaN         -1.0           NaN   \n",
       "2           1.0          NaN  ...        NaN          NaN           NaN   \n",
       "4           NaN          NaN  ...        NaN          NaN           0.0   \n",
       "5           NaN          NaN  ...        NaN          NaN           NaN   \n",
       "7           NaN          NaN  ...        NaN          NaN           NaN   \n",
       "\n",
       "   Pleural Effusion  Pleural Other  Fracture  Support Devices  Unnamed: 0  \\\n",
       "1              -1.0            NaN       1.0              NaN         0.0   \n",
       "2               NaN            NaN       1.0              NaN         1.0   \n",
       "4               NaN            NaN       NaN              NaN         2.0   \n",
       "5               0.0            NaN       NaN              NaN         3.0   \n",
       "7               0.0            NaN       NaN              1.0         4.0   \n",
       "\n",
       "   split    patient_id  \n",
       "1  train  patient00002  \n",
       "2  train  patient00002  \n",
       "4  train  patient00003  \n",
       "5  train  patient00004  \n",
       "7  train  patient00005  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_data_split = pd.read_csv(data_dir + 'chexpert_split_2021_08_20.csv').set_index('index')\n",
    "\n",
    "df_img_data = pd.read_csv(data_dir + 'train.csv')\n",
    "\n",
    "df_img_data = pd.concat([df_img_data,df_data_split], axis=1)\n",
    "df_img_data = df_img_data[~df_img_data.split.isna()]\n",
    "\n",
    "split =  df_img_data.Path.str.split(\"/\", expand = True)\n",
    "df_img_data[\"patient_id\"] = split[2]\n",
    "df_img_data = df_img_data.rename(columns={'Age': 'age'})\n",
    "df_img_data = df_img_data.rename(columns={'Sex': 'sex'})\n",
    "df_img_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient_id</th>\n",
       "      <th>race</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>Path</th>\n",
       "      <th>sex</th>\n",
       "      <th>age</th>\n",
       "      <th>Frontal/Lateral</th>\n",
       "      <th>AP/PA</th>\n",
       "      <th>No Finding</th>\n",
       "      <th>Enlarged Cardiomediastinum</th>\n",
       "      <th>...</th>\n",
       "      <th>Consolidation</th>\n",
       "      <th>Pneumonia</th>\n",
       "      <th>Atelectasis</th>\n",
       "      <th>Pneumothorax</th>\n",
       "      <th>Pleural Effusion</th>\n",
       "      <th>Pleural Other</th>\n",
       "      <th>Fracture</th>\n",
       "      <th>Support Devices</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>split</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "      <td>CheXpert-v1.0/train/patient24428/study22/view1...</td>\n",
       "      <td>Male</td>\n",
       "      <td>59</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>53688.0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "      <td>CheXpert-v1.0/train/patient24428/study39/view1...</td>\n",
       "      <td>Male</td>\n",
       "      <td>61</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>PA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>53689.0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "      <td>CheXpert-v1.0/train/patient24428/study61/view1...</td>\n",
       "      <td>Male</td>\n",
       "      <td>61</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>53690.0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "      <td>CheXpert-v1.0/train/patient24428/study35/view1...</td>\n",
       "      <td>Male</td>\n",
       "      <td>60</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>53691.0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>patient24428</td>\n",
       "      <td>White</td>\n",
       "      <td>Non-Hispanic/Non-Latino</td>\n",
       "      <td>CheXpert-v1.0/train/patient24428/study45/view1...</td>\n",
       "      <td>Male</td>\n",
       "      <td>61</td>\n",
       "      <td>Frontal</td>\n",
       "      <td>AP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>53692.0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     patient_id   race                ethnicity  \\\n",
       "0  patient24428  White  Non-Hispanic/Non-Latino   \n",
       "1  patient24428  White  Non-Hispanic/Non-Latino   \n",
       "2  patient24428  White  Non-Hispanic/Non-Latino   \n",
       "3  patient24428  White  Non-Hispanic/Non-Latino   \n",
       "4  patient24428  White  Non-Hispanic/Non-Latino   \n",
       "\n",
       "                                                Path   sex  age  \\\n",
       "0  CheXpert-v1.0/train/patient24428/study22/view1...  Male   59   \n",
       "1  CheXpert-v1.0/train/patient24428/study39/view1...  Male   61   \n",
       "2  CheXpert-v1.0/train/patient24428/study61/view1...  Male   61   \n",
       "3  CheXpert-v1.0/train/patient24428/study35/view1...  Male   60   \n",
       "4  CheXpert-v1.0/train/patient24428/study45/view1...  Male   61   \n",
       "\n",
       "  Frontal/Lateral AP/PA  No Finding  Enlarged Cardiomediastinum  ...  \\\n",
       "0         Frontal    AP         NaN                         NaN  ...   \n",
       "1         Frontal    PA         NaN                         NaN  ...   \n",
       "2         Frontal    AP         NaN                         NaN  ...   \n",
       "3         Frontal    AP         NaN                         NaN  ...   \n",
       "4         Frontal    AP         NaN                         NaN  ...   \n",
       "\n",
       "   Consolidation  Pneumonia  Atelectasis  Pneumothorax  Pleural Effusion  \\\n",
       "0            NaN        NaN          NaN           NaN               NaN   \n",
       "1            0.0        NaN          NaN           NaN              -1.0   \n",
       "2            NaN        NaN          NaN           NaN               1.0   \n",
       "3            NaN        NaN          1.0           NaN               NaN   \n",
       "4            NaN        NaN          NaN           NaN               1.0   \n",
       "\n",
       "   Pleural Other  Fracture  Support Devices  Unnamed: 0  split  \n",
       "0            NaN       NaN              1.0     53688.0  train  \n",
       "1            NaN       NaN              NaN     53689.0  train  \n",
       "2            NaN       NaN              1.0     53690.0  train  \n",
       "3            NaN       NaN              1.0     53691.0  train  \n",
       "4            NaN       NaN              1.0     53692.0  train  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_cxr = df_demo.merge(df_img_data, on=\"patient_id\")\n",
    "df_cxr.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "white = 'White'\n",
    "asian = 'Asian'\n",
    "black = 'Black'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['White', 'Asian', 'Black'], dtype=object)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = (df_cxr.race.str.contains(\"Black\", na=False))\n",
    "df_cxr.loc[mask, \"race\"] = black\n",
    "\n",
    "mask = (df_cxr.race.str.contains(\"White\", na=False))\n",
    "df_cxr.loc[mask, \"race\"] = white\n",
    "\n",
    "mask = (df_cxr.race.str.contains(\"Asian\", na=False))\n",
    "df_cxr.loc[mask, \"race\"] = asian\n",
    "\n",
    "df_cxr['race'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr = df_cxr[df_cxr.race.isin([asian,black,white])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr = df_cxr[df_cxr.ethnicity.isin([\"Non-Hispanic/Non-Latino\",\"Not Hispanic\"])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr = df_cxr[df_cxr[\"Frontal/Lateral\"]==\"Frontal\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr['race_label'] = df_cxr['race']\n",
    "\n",
    "df_cxr.loc[df_cxr['race_label'] == white, 'race_label'] = 0\n",
    "df_cxr.loc[df_cxr['race_label'] == asian, 'race_label'] = 1\n",
    "df_cxr.loc[df_cxr['race_label'] == black, 'race_label'] = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr['sex_label'] = df_cxr['sex']\n",
    "\n",
    "df_cxr.loc[df_cxr['sex_label'] == 'Male', 'sex_label'] = 0\n",
    "df_cxr.loc[df_cxr['sex_label'] == 'Female', 'sex_label'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = [\n",
    "    'No Finding',\n",
    "    'Enlarged Cardiomediastinum',\n",
    "    'Cardiomegaly',\n",
    "    'Lung Opacity',\n",
    "    'Lung Lesion',\n",
    "    'Edema',\n",
    "    'Consolidation',\n",
    "    'Pneumonia',\n",
    "    'Atelectasis',\n",
    "    'Pneumothorax',\n",
    "    'Pleural Effusion',\n",
    "    'Pleural Other',\n",
    "    'Fracture',\n",
    "    'Support Devices']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr['disease'] = df_cxr[labels[0]]\n",
    "df_cxr.loc[df_cxr[labels[0]] == 1, 'disease'] = labels[0]\n",
    "df_cxr.loc[df_cxr[labels[10]] == 1, 'disease'] = labels[10]\n",
    "df_cxr.loc[df_cxr['disease'].isna(), 'disease'] = 'Other'\n",
    "\n",
    "df_cxr['disease_label'] = df_cxr['disease']\n",
    "df_cxr.loc[df_cxr['disease_label'] == labels[0], 'disease_label'] = 0\n",
    "df_cxr.loc[df_cxr['disease_label'] == labels[10], 'disease_label'] = 1\n",
    "df_cxr.loc[df_cxr['disease_label'] == 'Other', 'disease_label'] = 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pre-processing of imaging dataset (this may take a few hours, needed only once)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# point to the parent directory that contains the folder 'CheXpert-v1.0'\n",
    "img_data_dir = '<path_to_imaging_data>'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from skimage.io import imread\n",
    "from skimage.io import imsave\n",
    "from skimage.transform import resize\n",
    "\n",
    "df_cxr['path_preproc'] = df_cxr['Path']\n",
    "\n",
    "preproc_dir = 'preproc_224x224/'\n",
    "out_dir = img_data_dir\n",
    "\n",
    "if not os.path.exists(out_dir + preproc_dir):\n",
    "    os.makedirs(out_dir + preproc_dir)\n",
    "\n",
    "for idx, p in enumerate(tqdm(df_cxr['Path'])):\n",
    "\n",
    "    split =  p.split(\"/\")\n",
    "    preproc_filename = split[2] + '_' + split[3] + '_' + split[4]\n",
    "    df_cxr.loc[idx, 'path_preproc'] = preproc_dir + preproc_filename\n",
    "    out_path = out_dir + preproc_dir + preproc_filename\n",
    "    \n",
    "    if not os.path.exists(out_path):\n",
    "        image = imread(img_data_dir + p)\n",
    "        image = resize(image, output_shape=(224, 224), preserve_range=True)\n",
    "        imsave(out_path, image.astype(np.uint8))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save full sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr.to_csv(data_dir + 'chexpert.sample.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating subsets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr = pd.read_csv(data_dir + 'chexpert.sample.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = df_cxr[df_cxr.split==\"train\"]\n",
    "df_val = df_cxr[df_cxr.split==\"validate\"]\n",
    "df_test = df_cxr[df_cxr.split==\"test\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.to_csv(data_dir + 'chexpert.sample.train.csv')\n",
    "df_val.to_csv(data_dir + 'chexpert.sample.val.csv')\n",
    "df_test.to_csv(data_dir + 'chexpert.sample.test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train_white = df_train[df_train['race'] == white]\n",
    "df_val_white = df_val[df_val['race'] == white]\n",
    "df_train_white.to_csv(data_dir + 'chexpert.sample.train.white.csv')\n",
    "df_val_white.to_csv(data_dir + 'chexpert.sample.val.white.csv')\n",
    "\n",
    "df_cxr_white = df_cxr[df_cxr['race'] == white]\n",
    "df_cxr_asian = df_cxr[df_cxr['race'] == asian]\n",
    "df_cxr_black = df_cxr[df_cxr['race'] == black]\n",
    "\n",
    "df_test_white = df_test[df_test['race'] == white]\n",
    "df_test_asian = df_test[df_test['race'] == asian]\n",
    "df_test_black = df_test[df_test['race'] == black]\n",
    "\n",
    "df_train_asian = df_train[df_train['race'] == asian]\n",
    "df_train_black = df_train[df_train['race'] == black]\n",
    "\n",
    "df_val_asian = df_val[df_val['race'] == asian]\n",
    "df_val_black = df_val[df_val['race'] == black]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train_male = df_train[df_train['sex'] == 'Male']\n",
    "df_val_male = df_val[df_val['sex'] == 'Male']\n",
    "df_train_male.to_csv(data_dir + 'chexpert.sample.train.male.csv')\n",
    "df_val_male.to_csv(data_dir + 'chexpert.sample.val.male.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cxr.loc[df_cxr['disease'] == 'Pleural Effusion', 'disease'] = 'Pleur. Eff.'\n",
    "df_cxr.loc[df_cxr['disease'] == 'No Finding', 'disease'] = 'No Find.'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 425.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fontsize = 16\n",
    "plt.rc('xtick', labelsize=fontsize) \n",
    "plt.rc('ytick', labelsize=fontsize)\n",
    "fig = sns.catplot(x='sex', y='age', hue='race', order=['Male', 'Female'], hue_order=['White', 'Asian', 'Black'], kind=\"box\", data=df_cxr)\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Age (years)', fontsize=fontsize)\n",
    "plt.ylim([0, 100])\n",
    "plt.legend(fontsize=fontsize, loc='lower center', ncol=3)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-age-sex-race.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 425.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='disease', y='age', hue='race', order=['Other', 'Pleur. Eff.', 'No Find.'], hue_order=['White', 'Asian', 'Black'], kind=\"box\", data=df_cxr)\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Age (years)', fontsize=fontsize)\n",
    "plt.ylim([0, 100])\n",
    "plt.legend(fontsize=fontsize, loc='lower center', ncol=3)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-age-pathology-race.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='disease', hue='sex', order=['Other', 'Pleur. Eff.', 'No Find.'], hue_order=['Male', 'Female'], data=df_cxr, kind='count')\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-sex-pathology.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 425.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='disease', hue='race', order=['Other', 'Pleur. Eff.', 'No Find.'], hue_order=['White', 'Asian', 'Black'], data=df_cxr, kind='count')\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-race-pathology.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAFYCAYAAACcQ30BAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAgu0lEQVR4nO3deZhcVZ3/8fcnYUkQkAABhEzYGnWCCGLLiEsC6AgiRAVFHRwFBoKgRMBBB2UTAi4oYIuMoAgqsjyIyPJTdgjooNIEUIJiGtl6ZAlJwIQkbPnOH+fWj6Ko7lSnq+rW6Xxez1NPdd17qurbVdWfPnXuufcqIjAzs7yMKrsAMzMbOoe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmG2h7ekiZI+q6k2yUtlhSSNqvTboykUyU9JmlJ0X5ynXajJB0t6SFJSyXdI2nvtvwyZmYlKaPn3QXsAywAbhuk3bnAQcBxwB7AY8C1kraraXcScAJwJvB+4HfApZJ2b2rVZmYdRO3eSUfSqIhYVvx8IPADYPOIeKiqzbbA3cABEXFesWwVYDZwf0RMLZZtADwKfD0ijq+6/43A+Ih4c1t+KTOzNmt7z7sS3MsxFXgBuKTqfi8CFwO7Slq9WLwrsBpwQc39LwC2kbT58Cs2M+s8nbrBcmvgwYhYXLN8Nimsu6raPQf01WkHMKllFZqZlWiVsgsYwLqkMfFa86vWV66fjleP/dS2ewVJ04BpAJMmTXrr7Nmz6zUzM+sEqrewU3veLRUR50REd0R0jx07tuxyzMyGrFPDewEwrs7ySk96flW7dSTV/meqbWdmNqJ0anjPBjaXtEbN8knA87w8xj0bWB3Ysk47gPtaVqGZWYk6NbyvAlYFPlpZUEwV/BhwXUQ8Vyy+hjQrZd+a+38SuDciHmxDrWZmbVfKBktJHyl+fGtx/X5Jc4G5ETEzIu6SdAlwhqRVgQeBQ4DNqQrqiHhS0mnA0ZIWArNIAb8LabqhmdmIVNZsk0trbp9VXM8Edip+3h84GZgBrAPcA+wWEbNq7vsVYBHweWAj4H5gn4i4uulVm5l1iLbvYdlpuru7o7e3t+wyzMwG4qmCZmYjhcPbzCxDDm8zsww5vM3MMuTwNjPLkMPbzCxDDm8zsww5vM3MMuTwNjPLkMPbzCxDDm8zsww5vM3MMuTwNjPLkMPbzCxDDm8zsww5vM3MMuTwNjPLkMPbzCxDDm8zsww5vM3MMuTwNjPLkMPbzCxDq5RdwMqmp6eHvr6+pj1ef38/ABMmTGjaY3Z1dTF9+vSmPV6n8nvROfxeDJ3DO3NLliwpuwQr+L3oHCvDe6GIKLuGUnV3d0dvb2/ZZaywSk+gp6en5ErM70XnGGHvheot9Ji3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZahjw1vSOyVdJ+lJSQslzZJ0QE2bMZJOlfSYpCWSbpc0uayazczapSPDW9KbgRuAVYGDgL2AO4BzJR1S1fTcYv1xwB7AY8C1krZra8FmZm22StkFDODjwGhgz4hYVCy7vgj1TwH/LWlb4N+AAyLiPABJM4HZwInA1PaXbWbWHh3Z8wZWA14AltQsf4aXa55atLmksjIiXgQuBnaVtHob6jQzK0Wnhvf5xXWPpI0lrSPpIOA9wOnFuq2BByNicc19Z5PCv6stlZqZlaAjh00i4l5JOwGXA4cWi18APhMRFxe31wUW1Ln7/Kr1dUmaBkwDmDhxYhMqNjNrr47seUvaCriM1IveE3gv8H3g+5L2He7jR8Q5EdEdEd3jx48f7sOZmbVdR/a8gVNIPe09IuKFYtmNktYDviPpIlKve9M69630uOfXWWdmNiJ0ZM8b2Aa4pyq4K/4ArAdsQOqVby5pjZo2k4Dngb6WV2lmVpJODe/Hge0krVaz/F+ApaRe9VWkeeAfrayUtArwMeC6iHiuTbWambVdpw6bnAlcClwl6SzSlMGpwCeA0yPieeAuSZcAZ0haFXgQOATYHBj2uLh1pp6eHvr6OvNL1Zw5cwCYPn16yZUMrKurq6Prs8Z1ZHhHxM8l7Q58CfghMAZ4APgscHZV0/2Bk4EZwDrAPcBuETGrrQVb2/T19fHXe2cxcc2Xyi7lVVZ7IX2RXfrQHSVXUt8ji0aXXYI1UUeGN0BE/Br49XLaLAGOLC62kpi45ksc071o+Q3tFWb0rll2CdZEnTrmbWZmg3B4m5llyOFtZpahjh3z7iSe4TA8nuFg1nwO7wb09fVx15/uY9kaAx4upTR6PgC484HHS66kvlGLvaOrWSs4vBu0bI11WTppj7LLyM6Y+64uuwSzEclj3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYZWKbsAM8tTT08PfX19ZZdR15w5cwCYPn16yZUMrKura1j1ObzNbIX09fVx1+y7YJ2yK6ljWbq663/vKreOgTw9/IdweJvZilsHlu20rOwqsjPqluGPWDu8LSv9/f08u3A0M3rXLLuU7Dy8cDSv6e8vuwxrEm+wNDPLkHvelpUJEyaw9MXHOKZ7UdmlZGdG75qMmTCh7DKsSdzzNjPLkMPbzCxDDm8zsww1NOYtaXVgR+DtwMbAWOAp4H7g1oj4W8sq7AD9/f2MWvwMY+67uuxSsjNq8Tz6+18suwyzEWfQ8JbUBRwO7Au8ljT1/RlgCbAuMAYISXcCZwE/iQhP+jQza7EBw1vS94CDgLuAE4FbgXsi4sWqNhuSeuMfAE4DviRpv4j4fUurbrMJEybwxHOrsHTSHmWXkp0x913NhAkblV2G2YgzWM97Y2CHiLh7oAYR8QRwBXCFpMOAg4FtgREV3mZmnWbA8I6IDw/lgSLiOaBn2BWZmdlyebaJmVmGGgpvSR+UtH/V7U0l3S5poaSfS/KBJszM2qjRnvcxwPiq26cBE4BzgMnACc0ty8zMBtNoeG8J/BFA0lhgd+DIiPgC8GVgSOPjQyFpd0m3Slok6R+SeiXtUrV+nKQfSnpK0rOSbpC0TavqMTPrBI2G9xjS3G6Ad5A2dF5X3L6fNDOl6SQdTJrNcifpH8RHgUuBNYr1Aq4CdgMOA/YGVgVuluQj8JjZiNXoUQUfAt4FzAQ+CNwZEc8U6zYg7bjTVJI2A84AjoqIM6pWXVv181TgncAuEXFzcb/bgQeBLwKdew4kM7NhaLTnfTZwgqRe4FDg3Kp1OwL3Nbsw4ADSHp3fH6TNVODvleAGKP6pXEX6J2NmNiI1FN4R8R3g08DtwAER8YOq1WsB57WgtncBfwE+LukBSS9K6pP02ao2WwP31rnvbGCiZ8GY2Ui13GETSasBhwA3RsSFtesj4uBWFEYaR98YOJW0UfQB0pj3mZJWKf6hrEsa0qk1v7geB7zqqP2SpgHTACZOnNj0ws3MWm25Pe+IeB74Oiko22kUqVd/cET8ICJuiohDgGuAo4uNlSskIs6JiO6I6B4/fvzy72Bm1mEaHfP+M7BFKwupY15xfX3N8uuADYHXAQtIvetalX80C1pTmplZuRoN7+OAY9s8f3r2ctYvK9psXWfdJOCRiPCJDs1sRGo0vL8ErAncVWw0vK3YcaZymdmC2i4vrnetWb4b0B8RjwNXAptImlJZKWltYM9inZnZiNToPO+XaM10wMH8CrgZOFvS+sDfSBss3wdUjrNyJWkGzAWSjiINkxwNCPhmm+s1M2ubhsI7InZqcR31njMkfQj4GvBV0tj2X4B9K7NeImKZpD2Ab5HO5DOGFOY7R8Sj7a7ZzKxdGu15lyIi/gF8trgM1GY+aYeeA9pVl5lZ2YYU3pLGAVuRerivEBG3NqsoMzMbXKNnjx8D/AjYhzSeXM/oZhVlZmaDa3S2ybHATqRd5AV8DjgQ+A1pz0efmdfMrI0aHTbZm3QG+YuBnwK/j4hZwHmSLiVN3/t1a0o0s07U398Pz8CoW3w2xSF7Gvqjf1gP0eirPhGYHREvAS8Ar6la9yPgY8OqwszMhqTRnvc80k46AI8C2wK3FbfXB8Y2uS4z63ATJkxgruaybKdlZZeSnVG3jGLCJsM7X0yj4f074C2koZHLgJMkrQW8CHyBNPZt1haPLBrNjN7OO9rvE4vTF9kN1+jMMHtk0WheX3YR1jSNhvc3SEMnADOALtIY+GhSsB/S/NLMXq2rq6vsEgb0/Jw5AIzZbKuSK6nv9XT262dD0+gelr1Ab/HzQmBvSasDqxc70pi1xfTpnXtmu0ptPT09JVdiK4MV3sMyIp4DnmtiLWZm1qCG5/hIeoukX0h6qjgl2fbF8lMk7da6Es3MrFZD4S3pXaQDPr0RuLDmfsuAzzS/NDMzG0ijPe+vA9eSTnxwZM26WcD2zSzKzMwG1+iY9/bAXsVhWqNm3VOATwRpZtZGjfa8lwJrDLDudcAzzSnHzMwa0Wh4/wY4XFL1kQMrPfD/AG5qalVmZjaoRodNjgV+C9wD/JwU3J+WdBrwVuBtrSmvc4xaPJ8x911ddhmvoqVpmn2MWbvkSuobtXg+sFHZZZiNOI3upHOPpMnAqcBXePmwsLcBUyLi/taVWL5O3ittzpyFAGy1ZacG5EYd/fqZ5arhnXSKQ8C+pzgxw7rA0xGxuGWVdRDv1WdmnabRed7vq/wcEUsj4u8rS3CbmXWiRjdYXiOpT9JRktZvaUVmZrZcjYb3LsAdwElAv6QLJU1pXVlmZjaYhsI7Im6JiE8AE0gzT7qBmyX9WdLni7PKm5lZmwzp5HMR8VREnBoRrwf+lbR35Wmk3vj5krZpRZFmZvZKK3TmUEm7A9OBtwNPkk5KPAWYJcknZjAza7GhHBJ2I0lfkfQgcDWwDvBJ4J8i4jOks+ucDRzXikLNzOxlDc3zlnQZsAfpGCcXAGdFxOzqNhHxkqQLgUObXqWZmb1CozvpbAUcDvw0IhYN0u5PwM7DLcrMzAbX6O7xb26w3UJg5rAqMjOz5RpwzLvYDX7IVvR+ZmbWuME2WD4k6QhJ6zTyQJLeIelK4KimVGZmZgMabNjkUOAU4BRJ15COIHgPMJd01vhxwBbADqSNmROB84BzWlmwmZkNEt4R8QtJVwAfIp1wYQYwhpdPwgDp0LAPA5cA50TE31pXqpmZVQy6wTIiXgIuAy6TtBqwHbAxKcTnAX+JiEdbXaSZmb3SUI7n/TzwhxbWYmZmDVqh3ePNzKxcDm8zsww5vM3MMuTwNjPLkMPbzCxDQwpvSaMkvUnSFEmvaVVRZmY2uKEcz/uzwOOkvSxvAt5QLP+lpOmtKc/MzOppKLwlHQR8B/gl8DHSnpUVtwF7N70yMzMbUKM97yOBb0fENODymnV/oeiFm5lZezQa3psD1w6w7lnSKdHMzKxNGg3vp4DNBlj3BuB/m1KNmZk1pNFjm1wNHCfpFtJRBAFC0vrAEaSxcDNb2TwNo27pwBnHlZM1rllqFQN7GthkeA/RaHgfQzo35b3A70mHhe0B3gg8CZw4vDLMLDddXV1llzCgOXPmALDVJluVXMkANhn+69foOSyfktRNOgnxrsADxX3PBE6PiH8Mq4oGFCeE2BU4OSKOqVo+DjiVdNzxscDtwBER8adW12S2Mps+vXNnCFdq6+npKbmS1hnKIWEXAicVl7aS9Alg2zrLBVxFGo8/DFgAHA3cLGm7iOhvZ51mZu3SgYNVr1T0rE8nTVesNRV4J/DvEXFRRFxTLBsFfLF9VZqZtVdDPW9JNw2yehnwDHAncG5EPNGMwqp8A7g3Ii6SdGHNuqnA3yPi5sqCiHhG0lXAB4HO/V5nZjYMjfa8RZoSuBOwKek0aJsWt/+ZNA/8WOBeSZOaVZykdwGfAj47QJOtSRtRa80GJkrq1G3NZmbD0mh4nwYsBbojYsuIeEdEbAm8rVj+VWAr0pnlT25GYcU5M88GvhUR9w/QbF3SOHet+cX1uAEee5qkXkm9c+fOHX6xZmZt1mh4zwBOiIhZ1Qsj4k5ScM8oNg6eCkxuUm1fJM0eaco/g2oRcU5EdEdE9/jx45v98GZmLdfobJPXk3rV9cwFKhMWHwCGfahYSROBrwAHAqtLWr1q9eqS1gEWknrd9XrX6xbX9XrlZmbZa7Tn/RBw0ADrphXrAdYH5g2vJAC2II2rX0AK4MoF4D+Ln7chjW1vXef+k4BHImJRnXVmZtlrtOd9InCBpD8Cl5H2qtyAdCjYNwH/VrR7L2kPzOG6m7RHZ62bSYF+LtAHXAnsL2lKRMwEkLQ2sCdQOzPFzGzEaHQPy4skPUUa3/4ysCrwAtALvC8ibiiaHgm8NNyiIuJp4Jba5WmfHB6OiFuK21eS9qi8QNJRvLyTjoBvDrcOM7NONZQ9LK8Hrpc0ijQ88lRELKtps7TJ9S2vpmWS9gC+BZxFGmq5Hdg5Ih5tZy1mZu3UcHhXFIH9ZOW2pCnApyPigGYWNsBzq86y+cABxcXMbKWwQrvHS+qSdKKkB0nj0Ps0tywzMxvMUE5A/Npi55bfAveTpvItAA4BNm5RfWZmVseg4S1plKTdJV0CPAZ8n7Rb/PeKJodHxNntOCSsmZm9bMAxb0nfJk0B3IC0C/zlwI+BG4C1gc+1o0AzM3u1wTZYHkE6Y86vgP0i4v/vfCMpWl2YmZkNbLBhk3NJu6B/ALhf0pmSdmhPWWZmNpgBwzsiDgI2AvYl7YxzMHC7pD8DXyL1ys3MrASDbrCMiKXFGWp2AyaS9l58Cfgv0l6MX5f0SUljWl+qmZlVNDxVMCIei4hvRsSbgB1IM062An5CmoliZmZtskI76UREb0QcRprfvTd1jkNiZmatM+Td46tFxAukKYSXN6ccMzNrRMefPd7MzF7N4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZcnibmWXI4W1mliGHt5lZhhzeZmYZ6tjwlvQRSZdJeljSEkn3S/qapLVq2o2T9ENJT0l6VtINkrYpq24zs3ZYpewCBvGfwCPAl4F+4C3ACcDOkt4REcskCbgK2Aw4DFgAHA3cLGm7iOgvo3DLQ09PD319fU17vLvvvhuAyZMns9122zXlMbu6upg+fXpTHstGlk4O7z0jYm7V7ZmS5gM/BnYCbgKmAu8EdomImwEk3Q48CHwR8KfezEakjg3vmuCuuKO43qS4ngr8vRLcxf2ekXQV8EEc3jaIZvZoJ0+e/Irbd999N7feemvTHt+sVseOeQ9gSnH95+J6a+DeOu1mAxMlrdmWqszM2iyb8Ja0CXAicENE9BaL1yWNc9eaX1yPG+CxpknqldQ7d269Dr6ZWWfLIryLHvQVwIvA/sN9vIg4JyK6I6J7/Pjxw67PzKzdOnbMu0LSWNKMki2AKTUzSBZQv3e9btV6M7MRp6N73pJWBX4OdAO7R8SfaprMJo1715oEPBIRi1pcoplZKTo2vCWNAn4G7AJ8KCJ+V6fZlcAmkqZU3W9tYM9inZnZiNTJwybfAz4KnAw8K+ntVev6i+GTK4HbgQskHcXLO+kI+Gab621Is3cMmTNnDtDcaW/eMcTazX8XQ9exPW/g/cX1V0gBXX05ECAilgF7ANcDZwGXAy8BO0fEo+0uuAxjx45l7NixZZdh1lFWhr8LRUTZNZSqu7s7ent7l9/QbBC1O+kA3knHmkX1FnZyz9vMzAbg8DZrgtpetnvd1moObzOzDHXybBOzrLi3be3knreZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpYhh7eZWYYc3mZmGXJ4m5llyOFtZpahERHekv5J0s8lPSPpH5J+IWli2XWZmbVK9uEtaQ3gJuCNwKeBfwe2Am6W9JoyazMza5VVyi6gCQ4CtgDeEBF9AJL+CMwBDgZOK7E2M7OWyL7nDUwFflcJboCIeBD4LfDB0qoyM2uhkRDeWwP31lk+G5jU5lrMzNpiJAybrAssqLN8PjCu3h0kTQOmFTcXSbq/RbW1y/rAU2UXYYDfi04yUt6LayJit9qFIyG8hywizgHOKbuOZpHUGxHdZddhfi86yUh/L0bCsMkC6vewB+qRm5llbySE92zSuHetScB9ba7FzKwtRkJ4Xwm8XdIWlQWSNgPeWaxbGYyYIaARwO9F5xjR74UiouwahqXYEeceYAlwDBDAScBawJsjYlGJ5ZmZtUT2Pe+IeBbYBfgr8FPgZ8CDwC4ObjMbqbLveZuZrYyy73nnStL7JP1a0jxJSyX9VdI3JI2rarOOpBMkbV/n/rdI+k17q+48kvaTFFWXhZLukfQ5SatUtXtI0vklljooSefX/B7Vl1/WtP2ypEckvSjp7mLZRpKulDS/uM/hJfwar1L1/jxd/dku1q1SrDuhyc9V7/LeqvWbNeP5iucs7XO1Us7zLpukLwMnA78EDiTtUPRW4EvA3pJ2johHgXWA44F+YFYpxebjo6TXae3i5+8CGwDHlVnUEM0lHe6h1vzKD5J2IH12TiV9fhYWq44DpgD7AY8BD7WuzBXyWtLn+7/a8FyVz0K1+0jbxnYkvT7Zc3i3maSdgRnAGRFxRNWqmZIuB+4EfgLsXEZ91SStHhHPlV1Hg+6uOr7NdZK6gM9TYnivwOv3fET8bjlt/rm4/n5E/K1m+T0RcfmQimyf64DDJJ0eEU+0+LmqPwu15rb4udvGwybt90VST+ro2hXFAbW+Duwk6V9IG14BflD19W+/6vsUXwdnSVos6V5JH659XEnbFl+pF0haIum3kt5d0+Z8Sf2SdpT0P5KWAN9szq9cijuAtSVtMFADSZtL+pmkuZKek3R37etXvC4P1bnvLZJuqbq9U/H+7CXpB5LmAk0NqeL5zi9uPlA83/mSAtgJeHfV52SzZj53E8woro9ZXkNJO0i6QdIiSc9KurH4xjEs9YZNimGPCyR9XNKfi+frlfSuOvf/fNF+adHm3bVt2snh3UbFGOwU4PqIWDpAs8rc9N2AvYqfv0b6urcj8P+q2m4JfId02Nu9SF8HLy16nZXn3B74H9IepwcBewPzgBskvbXmuV8LXAxcBLwfuHDov2XH2Bx4Cag740jSPwG/B7YFjiANV8wCLpNUb+iiUd8FRDqu/H5DvXMxDlx7UbH6UNJnAdL7vSNpWG1H4I/AXbz8Oem0oYHHgDOBaZI2HaiRpDcDM0l7Te8HfIo0FDZT0rYNPtfomtdv9HLavxv4AnAs8DFgNHC1pHWq6voP4AzgZuBDpH+iFzHA8ZPawcMm7bUeMJbBxyMr6zYk/TEC/G2Ar9PrA5MjYg6ApFmkP5J9gFOKNqcCj5CmTj5ftLuWdCTGY0kfxIo1gU9GxBVD+aU6xOjin+NapN9/L+CqiFg8QPsTSCE7JSLmFcuuLUL9RFZ8B68/RMSBK3jfTYAX6iw/CvhWRNwnqTJUcldEPFT8/LCkhcCLDQy7lOkbpGPsHw8cMECb44DngPdExNMAkq4n/V0cz8sdmsH8peb2b4FX9aSrrA1sFxELiud7nPTNbXfgQkmjSJ+XayNi/8qdim9XFzdQT0s4vPM2pxLcABHxpKQngYkAksaSevqnAMtUNfsCuAHYt+bxXgCubm3JLVP9B7uMNN//8EHa7wb8Cnim5nW5FjhV0toR8Y8VqGM4Y85PAh+os/zRYTxmx4iI+ZK+DRwv6RvAA3WaTQaurgR3cb9/SLoS2LPBp/owr9xguXCghoXbK8Fd+FNxXTmV4oTicnzN/S4DXmywpqZzeLfXPGApsNkgbSrrGvmDnV9n2XPAmOLndUlfAY8tLq8iaVRELCtuzo2Ilxp43k5U+YNdCDw8yLBUxQakr+SfGmD9esCKhPdwhiteiIjeYdw/B6cDh5G+3dR2HiB9Zuu9ho/T+BDFvYNssKznFX9HEfFcMVJV+Tt6XXH9RE27FyXNoyQO7zYq3uyZwL9KGjNAwFTGW29qwlM+TeqFfo80g6VeTcuqbzbhOcsy1D/YecBtpK/y9fy9uF4KrFZn/XrFY9TK+TVsuYhYJOlrwLdJQ3q15gMb1Vm+EeUdJbTyz2TD6oXFN7b12l9O4vBuv28B15OGMo6sXiFpc9Jc2Fsj4veSKv/xx67IE0XEs5JuI22Um1UT1Cu7a0gb9mZHxJJB2j0MbChpfETMBZC0JfAG0oZgG7qzSJ/9GXXWzQR2l7RWRCwEkLQWacjklrZV+Er9pG/C+wA/qlq+NyVmqMO7zSLiBknHA18tpiz9hNSj2J60A8MzpJkKkL6mzQM+rnRS5WeBB6s2sDXiSOBW0sa4c0m9iPWL5xsdEe3YaaITHQf8AbhV0pmkDWLjgDcBW0REZYPapaQDnV0g6TTSa3c0DZ6hRdKNwKYR0bXcxrCapLfXWb44Iv7YyPNVPe8U4EbggIio+62rLMWwxInUP+rfScAewI3FuHiQOjRrkIZa2i4ilkn6KvBDSeeRNlJ2kf5eXzG0VsykeQA4MSJaWq+nCpageFPfD7wGOI+0A8OhpCDvjohHinbLSHtgjiNtYLyDxjfaVJ5rFvA20j+BnuK5vgNsQwr1lVLxGneT9ro7hfRt6L9JG3hvqmrXB3yENBPkl6R5+keSDoTWiNE03kkaD9xe57IiUzZVPHen/o2fB8ypXVj8k9qJFIo/Jh1sbhFpVtA97Sywpq5zSRvAdwGuAPYHPsGrh3La9rr7wFRmZhnq1P/KZmY2CIe3mVmGHN5mZhlyeJuZZcjhbWaWIYe3mVmGHN5mZhlyeJuZZej/AF8hqTnk8kHhAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='disease', y='age', order=['Other', 'Pleur. Eff.', 'No Find.'], kind=\"box\", data=df_cxr)\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Age (years)', fontsize=fontsize)\n",
    "plt.ylim([0, 100])\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-age-pathology.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='sex', y='age', order=['Male', 'Female'], kind=\"box\", data=df_cxr)\n",
    "sns.stripplot(x='sex', y='age', color='k', order=['Male', 'Female'], alpha=0.01, data=df_cxr, ax=fig.ax)\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Age (years)', fontsize=fontsize)\n",
    "plt.ylim([0, 100])\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-age-sex.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='race', y='age', order=['White', 'Asian', 'Black'], kind=\"box\", data=df_cxr)\n",
    "sns.stripplot(x='race', y='age', color='k', order=['White', 'Asian', 'Black'], alpha=0.01, data=df_cxr, ax=fig.ax)\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Age (years)', fontsize=fontsize)\n",
    "plt.ylim([0, 100])\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-age-race.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='race', order=['White', 'Asian', 'Black'], data=df_cxr, kind='count')\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-race.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='sex', order=['Male', 'Female'], data=df_cxr, kind='count')\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-sex.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='disease', order=['Other', 'Pleur. Eff.', 'No Find.'], data=df_cxr, kind='count')\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-pathology.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='race', hue='sex', order=['White', 'Asian', 'Black'], hue_order=['Male', 'Female'], data=df_cxr, kind='count')\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-sex-race.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 425.125x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = sns.catplot(x='sex', hue='race', order=['Male', 'Female'], hue_order=['White', 'Asian', 'Black'], data=df_cxr, kind='count')\n",
    "fig._legend.remove()\n",
    "plt.xlabel('', fontsize=fontsize)\n",
    "plt.ylabel('Count', fontsize=fontsize)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.show()\n",
    "# fig.savefig(\"CheXpert-race-sex.png\", bbox_inches='tight', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}