Downloads: 1
[d88862]:
/
3. Applying AI to EHR Data /
3. EHR Transformations & Feature Engineering /
Numerical Feature Column API Exercise /
notebook(2).ipynb
2351 lines (2350 with data), 105.5 kB
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Exercise L3 - 1: Diagnose Dataset Level and Select Last Encounter"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Instructions\n",
"- Given the dataset, convert the dataset to a longitudinal level but select only the last encounter for each patient.\n",
"- Assume that that the order of encounter IDs is indicative of the time for encounter. In other words a lower number encounter will come before a higher numbered encounter."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"ehr_level_dataset_path = \"./data/ehr_level_exercise_dataset.csv\""
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Level of Dataset\n",
"What level is the dataset at? Is at the line or encounter level? "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Solution"
]
},
{
"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>ENCOUNTER_ID</th>\n",
" <th>PATIENT_ID</th>\n",
" <th>PRINCIPAL_DIAGNOSIS_CODE</th>\n",
" <th>PROCEDURE_CODE</th>\n",
" <th>MEDICATION_CODE</th>\n",
" <th>LAB_CODE</th>\n",
" <th>LABEL</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>udacity_health_encounter_id_1</td>\n",
" <td>udacity_health_patient_id_186</td>\n",
" <td>dx_code_15406</td>\n",
" <td>['procedure_code_58552', 'procedure_code_39776...</td>\n",
" <td>['medication_code_2350', 'medication_code_8630...</td>\n",
" <td>['lab_code_8835', 'lab_code_9859', 'lab_code_9...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>udacity_health_encounter_id_10</td>\n",
" <td>udacity_health_patient_id_188</td>\n",
" <td>dx_code_74047</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_7789', 'medication_code_3560...</td>\n",
" <td>[]</td>\n",
" <td>[1, 1, 1, 1]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>udacity_health_encounter_id_100</td>\n",
" <td>udacity_health_patient_id_585</td>\n",
" <td>dx_code_71465</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_7982', 'medication_code_2452...</td>\n",
" <td>['lab_code_4198', 'lab_code_6603', 'lab_code_3...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>udacity_health_encounter_id_1000</td>\n",
" <td>udacity_health_patient_id_525</td>\n",
" <td>dx_code_61569</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_4036']</td>\n",
" <td>[]</td>\n",
" <td>[0]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>udacity_health_encounter_id_1001</td>\n",
" <td>udacity_health_patient_id_950</td>\n",
" <td>dx_code_90172</td>\n",
" <td>['procedure_code_30555']</td>\n",
" <td>['medication_code_6755', 'medication_code_5045']</td>\n",
" <td>['lab_code_9112']</td>\n",
" <td>[0, 0, 0, 0]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID PATIENT_ID \\\n",
"0 udacity_health_encounter_id_1 udacity_health_patient_id_186 \n",
"1 udacity_health_encounter_id_10 udacity_health_patient_id_188 \n",
"2 udacity_health_encounter_id_100 udacity_health_patient_id_585 \n",
"3 udacity_health_encounter_id_1000 udacity_health_patient_id_525 \n",
"4 udacity_health_encounter_id_1001 udacity_health_patient_id_950 \n",
"\n",
" PRINCIPAL_DIAGNOSIS_CODE PROCEDURE_CODE \\\n",
"0 dx_code_15406 ['procedure_code_58552', 'procedure_code_39776... \n",
"1 dx_code_74047 [] \n",
"2 dx_code_71465 [] \n",
"3 dx_code_61569 [] \n",
"4 dx_code_90172 ['procedure_code_30555'] \n",
"\n",
" MEDICATION_CODE \\\n",
"0 ['medication_code_2350', 'medication_code_8630... \n",
"1 ['medication_code_7789', 'medication_code_3560... \n",
"2 ['medication_code_7982', 'medication_code_2452... \n",
"3 ['medication_code_4036'] \n",
"4 ['medication_code_6755', 'medication_code_5045'] \n",
"\n",
" LAB_CODE \\\n",
"0 ['lab_code_8835', 'lab_code_9859', 'lab_code_9... \n",
"1 [] \n",
"2 ['lab_code_4198', 'lab_code_6603', 'lab_code_3... \n",
"3 [] \n",
"4 ['lab_code_9112'] \n",
"\n",
" LABEL \n",
"0 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... \n",
"1 [1, 1, 1, 1] \n",
"2 [0, 0, 0, 0, 0, 0, 0] \n",
"3 [0] \n",
"4 [0, 0, 0, 0] "
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"ehr_level_df = pd.read_csv(ehr_level_dataset_path)\n",
"ehr_level_df.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"**Tests**\n",
"- Line: Total number of rows > Number of Unique Encounters\n",
"- Encounter level: Total Number of Rows = Number of Unique Encounters\n"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dataset is not at the line level\n"
]
}
],
"source": [
"# Line Test\n",
"try:\n",
" assert len(ehr_level_df) > ehr_level_df['ENCOUNTER_ID'].nunique() \n",
" print(\"Dataset could be at the line level\")\n",
"except:\n",
" print(\"Dataset is not at the line level\")"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dataset could be at the encounter level\n"
]
}
],
"source": [
"# Encounter Test\n",
"try:\n",
" assert len(ehr_level_df) == ehr_level_df['ENCOUNTER_ID'].nunique()\n",
" print(\"Dataset could be at the encounter level\")\n",
"except:\n",
" print(\"Dataset is not at the encounter level\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"**Answer:** Dataset is at the encounter level and you can probably guess by seeing the arrays for the code sets but we did a few simple tests to confirm."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Select Last Encounter for each Patient"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So in many cases you may only want a snapshot of a patient's history for your modeling objective. In some cases it might be important to see the changes over time but in other cases you only want the most recent case or depending on the model the first case could also be used. Really important to know how the context for how the model will be deployed in production and the time state you will be getting data."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"# select last encounter for each patient\n",
"#convert encounter id column to a numerical value\n",
"def convert_encounter_id_to_number(df, encounter_id):\n",
" df[\"ENCOUNTER_ID_NUMBER\"] = df[encounter_id].str.replace('udacity_health_encounter_id_', '').astype(int)\n",
" return df\n",
"\n",
"def select_last_encounter(df, patient_id, encounter_id):\n",
" df = df.sort_values(encounter_id)\n",
" last_encounter_values = df.groupby(patient_id)[encounter_id].tail(1).values\n",
" return df[df[encounter_id].isin(last_encounter_values)] "
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"ehr_encounter_number_df = convert_encounter_id_to_number(ehr_level_df, \"ENCOUNTER_ID\")\n",
"last_encounter_df = select_last_encounter(ehr_encounter_number_df, \"PATIENT_ID\", \"ENCOUNTER_ID_NUMBER\" )"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"#take subset of output\n",
"test_last_encounter_df = last_encounter_df[['ENCOUNTER_ID', 'ENCOUNTER_ID_NUMBER', 'PATIENT_ID']]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Test cases \n",
"- PATIENT_IDS - udacity_health_patient_id_309, udacity_health_patient_id_418, udacity_health_patient_id_908"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>ENCOUNTER_ID</th>\n",
" <th>PATIENT_ID</th>\n",
" <th>PRINCIPAL_DIAGNOSIS_CODE</th>\n",
" <th>PROCEDURE_CODE</th>\n",
" <th>MEDICATION_CODE</th>\n",
" <th>LAB_CODE</th>\n",
" <th>LABEL</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>94</th>\n",
" <td>udacity_health_encounter_id_1091</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_62502</td>\n",
" <td>['procedure_code_69597']</td>\n",
" <td>['medication_code_9253', 'medication_code_7444...</td>\n",
" <td>['lab_code_9415', 'lab_code_8179']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>1091</td>\n",
" </tr>\n",
" <tr>\n",
" <th>450</th>\n",
" <td>udacity_health_encounter_id_1436</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_37825</td>\n",
" <td>['procedure_code_14044', 'procedure_code_47911...</td>\n",
" <td>['medication_code_4081', 'medication_code_6050...</td>\n",
" <td>['lab_code_6246', 'lab_code_7197']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>1436</td>\n",
" </tr>\n",
" <tr>\n",
" <th>605</th>\n",
" <td>udacity_health_encounter_id_1582</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_79663</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_4450', 'medication_code_4900...</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0]</td>\n",
" <td>1582</td>\n",
" </tr>\n",
" <tr>\n",
" <th>665</th>\n",
" <td>udacity_health_encounter_id_1643</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_60773</td>\n",
" <td>['procedure_code_39999', 'procedure_code_31842...</td>\n",
" <td>['medication_code_413', 'medication_code_5966'...</td>\n",
" <td>['lab_code_4295', 'lab_code_2383']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>1643</td>\n",
" </tr>\n",
" <tr>\n",
" <th>684</th>\n",
" <td>udacity_health_encounter_id_1662</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_29272</td>\n",
" <td>['procedure_code_12589', 'procedure_code_3005'...</td>\n",
" <td>['medication_code_9845', 'medication_code_6602...</td>\n",
" <td>['lab_code_4330', 'lab_code_3499', 'lab_code_1...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>1662</td>\n",
" </tr>\n",
" <tr>\n",
" <th>867</th>\n",
" <td>udacity_health_encounter_id_1837</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_56368</td>\n",
" <td>['procedure_code_14069', 'procedure_code_15230...</td>\n",
" <td>['medication_code_5820', 'medication_code_8028...</td>\n",
" <td>['lab_code_8964']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>1837</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1024</th>\n",
" <td>udacity_health_encounter_id_1985</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_38581</td>\n",
" <td>['procedure_code_1431', 'procedure_code_71061']</td>\n",
" <td>['medication_code_3134', 'medication_code_4868...</td>\n",
" <td>['lab_code_7702']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>1985</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1042</th>\n",
" <td>udacity_health_encounter_id_2000</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_91230</td>\n",
" <td>[]</td>\n",
" <td>[]</td>\n",
" <td>['lab_code_8062']</td>\n",
" <td>[0]</td>\n",
" <td>2000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1196</th>\n",
" <td>udacity_health_encounter_id_2142</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_58527</td>\n",
" <td>['procedure_code_56979', 'procedure_code_24797...</td>\n",
" <td>['medication_code_7631', 'medication_code_7508...</td>\n",
" <td>['lab_code_3908', 'lab_code_1500']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>2142</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1289</th>\n",
" <td>udacity_health_encounter_id_2229</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_46826</td>\n",
" <td>['procedure_code_50916', 'procedure_code_20032...</td>\n",
" <td>['medication_code_9204', 'medication_code_1282...</td>\n",
" <td>['lab_code_5607', 'lab_code_5139']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>2229</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1340</th>\n",
" <td>udacity_health_encounter_id_2281</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_39484</td>\n",
" <td>['procedure_code_33400', 'procedure_code_56', ...</td>\n",
" <td>['medication_code_1938', 'medication_code_4076...</td>\n",
" <td>['lab_code_9178', 'lab_code_3557', 'lab_code_2...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>2281</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1596</th>\n",
" <td>udacity_health_encounter_id_2531</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_75423</td>\n",
" <td>['procedure_code_15600', 'procedure_code_19008...</td>\n",
" <td>['medication_code_7265', 'medication_code_2247...</td>\n",
" <td>['lab_code_8336', 'lab_code_7292', 'lab_code_8...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>2531</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1702</th>\n",
" <td>udacity_health_encounter_id_2631</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_38210</td>\n",
" <td>['procedure_code_121']</td>\n",
" <td>['medication_code_7480', 'medication_code_7917...</td>\n",
" <td>['lab_code_4777']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>2631</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2069</th>\n",
" <td>udacity_health_encounter_id_2989</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_80145</td>\n",
" <td>['procedure_code_28893', 'procedure_code_9124'...</td>\n",
" <td>['medication_code_43', 'medication_code_5602',...</td>\n",
" <td>['lab_code_3392', 'lab_code_9665', 'lab_code_4...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>2989</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2291</th>\n",
" <td>udacity_health_encounter_id_3209</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_74710</td>\n",
" <td>['procedure_code_41970', 'procedure_code_64032...</td>\n",
" <td>['medication_code_7596', 'medication_code_7481...</td>\n",
" <td>['lab_code_9105']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>3209</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2507</th>\n",
" <td>udacity_health_encounter_id_3416</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_43343</td>\n",
" <td>['procedure_code_11834', 'procedure_code_12304...</td>\n",
" <td>['medication_code_3767', 'medication_code_7630...</td>\n",
" <td>['lab_code_774', 'lab_code_9591', 'lab_code_31...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>3416</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2696</th>\n",
" <td>udacity_health_encounter_id_360</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_94868</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_193', 'medication_code_3556'...</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0]</td>\n",
" <td>360</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2890</th>\n",
" <td>udacity_health_encounter_id_3792</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_37738</td>\n",
" <td>['procedure_code_11852', 'procedure_code_45278...</td>\n",
" <td>['medication_code_7882', 'medication_code_6562...</td>\n",
" <td>['lab_code_3787', 'lab_code_6452', 'lab_code_6...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>3792</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2960</th>\n",
" <td>udacity_health_encounter_id_3859</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_17144</td>\n",
" <td>['procedure_code_69917']</td>\n",
" <td>['medication_code_2152', 'medication_code_7078...</td>\n",
" <td>['lab_code_5682']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>3859</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3064</th>\n",
" <td>udacity_health_encounter_id_3961</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_61465</td>\n",
" <td>['procedure_code_15248', 'procedure_code_65724...</td>\n",
" <td>['medication_code_701', 'medication_code_8324'...</td>\n",
" <td>['lab_code_7691', 'lab_code_4598', 'lab_code_1...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>3961</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3167</th>\n",
" <td>udacity_health_encounter_id_4058</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_73636</td>\n",
" <td>['procedure_code_5646', 'procedure_code_1381',...</td>\n",
" <td>['medication_code_4201', 'medication_code_5328...</td>\n",
" <td>['lab_code_9083', 'lab_code_1392', 'lab_code_1...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>4058</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3280</th>\n",
" <td>udacity_health_encounter_id_4168</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_9795</td>\n",
" <td>['procedure_code_44804', 'procedure_code_10917...</td>\n",
" <td>['medication_code_2738', 'medication_code_9919...</td>\n",
" <td>['lab_code_4690', 'lab_code_5570', 'lab_code_2...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>4168</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3353</th>\n",
" <td>udacity_health_encounter_id_4239</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_28764</td>\n",
" <td>['procedure_code_31873', 'procedure_code_29930...</td>\n",
" <td>['medication_code_7607', 'medication_code_6643...</td>\n",
" <td>['lab_code_8492', 'lab_code_3309', 'lab_code_6...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>4239</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3561</th>\n",
" <td>udacity_health_encounter_id_4439</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_31393</td>\n",
" <td>['procedure_code_69203', 'procedure_code_53820...</td>\n",
" <td>['medication_code_9780', 'medication_code_7008...</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>4439</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3585</th>\n",
" <td>udacity_health_encounter_id_4462</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_53624</td>\n",
" <td>['procedure_code_64278', 'procedure_code_19230']</td>\n",
" <td>['medication_code_8759', 'medication_code_59',...</td>\n",
" <td>['lab_code_3790', 'lab_code_591', 'lab_code_89...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>4462</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3718</th>\n",
" <td>udacity_health_encounter_id_4589</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_67438</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_8103']</td>\n",
" <td>[]</td>\n",
" <td>[0]</td>\n",
" <td>4589</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3752</th>\n",
" <td>udacity_health_encounter_id_4619</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_79460</td>\n",
" <td>['procedure_code_8648', 'procedure_code_22388'...</td>\n",
" <td>['medication_code_4773', 'medication_code_5839...</td>\n",
" <td>['lab_code_5055']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>4619</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3916</th>\n",
" <td>udacity_health_encounter_id_478</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_43778</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_1646', 'medication_code_4346']</td>\n",
" <td>[]</td>\n",
" <td>[0, 0]</td>\n",
" <td>478</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4552</th>\n",
" <td>udacity_health_encounter_id_5398</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_58513</td>\n",
" <td>['procedure_code_3629', 'procedure_code_56509'...</td>\n",
" <td>['medication_code_5697', 'medication_code_9845...</td>\n",
" <td>['lab_code_4909']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>5398</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4687</th>\n",
" <td>udacity_health_encounter_id_553</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_88408</td>\n",
" <td>['procedure_code_19471', 'procedure_code_17060...</td>\n",
" <td>['medication_code_1541', 'medication_code_5218...</td>\n",
" <td>['lab_code_4271', 'lab_code_7946', 'lab_code_9...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>553</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4771</th>\n",
" <td>udacity_health_encounter_id_5611</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_46410</td>\n",
" <td>['procedure_code_71656', 'procedure_code_29912...</td>\n",
" <td>['medication_code_3365', 'medication_code_862'...</td>\n",
" <td>['lab_code_8280', 'lab_code_9803', 'lab_code_1...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>5611</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4785</th>\n",
" <td>udacity_health_encounter_id_5624</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_96267</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_1343']</td>\n",
" <td>['lab_code_9589']</td>\n",
" <td>[0, 0]</td>\n",
" <td>5624</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4796</th>\n",
" <td>udacity_health_encounter_id_5635</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_67422</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_2864']</td>\n",
" <td>[]</td>\n",
" <td>[0]</td>\n",
" <td>5635</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4862</th>\n",
" <td>udacity_health_encounter_id_5701</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_4007</td>\n",
" <td>['procedure_code_8263', 'procedure_code_7509']</td>\n",
" <td>['medication_code_3925', 'medication_code_3163']</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0, 0]</td>\n",
" <td>5701</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4881</th>\n",
" <td>udacity_health_encounter_id_5721</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_11147</td>\n",
" <td>['procedure_code_68127', 'procedure_code_3768'...</td>\n",
" <td>['medication_code_4722', 'medication_code_2681...</td>\n",
" <td>['lab_code_8853', 'lab_code_3111', 'lab_code_2...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>5721</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5132</th>\n",
" <td>udacity_health_encounter_id_597</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_25907</td>\n",
" <td>['procedure_code_41138', 'procedure_code_28596']</td>\n",
" <td>['medication_code_4777', 'medication_code_4318...</td>\n",
" <td>['lab_code_1043']</td>\n",
" <td>[0, 0, 0, 0, 0, 0]</td>\n",
" <td>597</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5306</th>\n",
" <td>udacity_health_encounter_id_6139</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_40784</td>\n",
" <td>['procedure_code_48973', 'procedure_code_7953'...</td>\n",
" <td>['medication_code_3191', 'medication_code_105'...</td>\n",
" <td>['lab_code_5078', 'lab_code_2361', 'lab_code_9...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>6139</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5385</th>\n",
" <td>udacity_health_encounter_id_622</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_25556</td>\n",
" <td>['procedure_code_10462']</td>\n",
" <td>['medication_code_3626', 'medication_code_311']</td>\n",
" <td>['lab_code_4960']</td>\n",
" <td>[0, 0, 0, 0]</td>\n",
" <td>622</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5472</th>\n",
" <td>udacity_health_encounter_id_6305</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_93595</td>\n",
" <td>['procedure_code_6005', 'procedure_code_13203'...</td>\n",
" <td>['medication_code_4513']</td>\n",
" <td>['lab_code_7349', 'lab_code_1043', 'lab_code_9...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>6305</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5564</th>\n",
" <td>udacity_health_encounter_id_6394</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_17875</td>\n",
" <td>['procedure_code_42424']</td>\n",
" <td>[]</td>\n",
" <td>['lab_code_5222']</td>\n",
" <td>[0, 0]</td>\n",
" <td>6394</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5669</th>\n",
" <td>udacity_health_encounter_id_6497</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_36095</td>\n",
" <td>['procedure_code_33881', 'procedure_code_6396'...</td>\n",
" <td>['medication_code_6849', 'medication_code_2364...</td>\n",
" <td>['lab_code_2570', 'lab_code_2383', 'lab_code_8...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>6497</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5684</th>\n",
" <td>udacity_health_encounter_id_6512</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_92784</td>\n",
" <td>['procedure_code_31090', 'procedure_code_28839...</td>\n",
" <td>['medication_code_5274', 'medication_code_6976...</td>\n",
" <td>['lab_code_4639']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>6512</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5829</th>\n",
" <td>udacity_health_encounter_id_6662</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_26279</td>\n",
" <td>['procedure_code_47253', 'procedure_code_52416...</td>\n",
" <td>['medication_code_2602', 'medication_code_3320...</td>\n",
" <td>['lab_code_5441']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>6662</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5846</th>\n",
" <td>udacity_health_encounter_id_668</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_22666</td>\n",
" <td>['procedure_code_10507', 'procedure_code_39636...</td>\n",
" <td>['medication_code_4676', 'medication_code_4760...</td>\n",
" <td>['lab_code_7518', 'lab_code_1774', 'lab_code_5...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>668</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5946</th>\n",
" <td>udacity_health_encounter_id_6777</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_78269</td>\n",
" <td>['procedure_code_58888', 'procedure_code_72745']</td>\n",
" <td>['medication_code_227', 'medication_code_7799']</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0, 0]</td>\n",
" <td>6777</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6508</th>\n",
" <td>udacity_health_encounter_id_7327</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_18813</td>\n",
" <td>['procedure_code_9329', 'procedure_code_28542'...</td>\n",
" <td>['medication_code_4006', 'medication_code_485'...</td>\n",
" <td>['lab_code_71', 'lab_code_7769', 'lab_code_659...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>7327</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6892</th>\n",
" <td>udacity_health_encounter_id_7696</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_24785</td>\n",
" <td>['procedure_code_31137', 'procedure_code_23811...</td>\n",
" <td>['medication_code_2626', 'medication_code_6310...</td>\n",
" <td>['lab_code_8317', 'lab_code_2116', 'lab_code_2...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>7696</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6969</th>\n",
" <td>udacity_health_encounter_id_7772</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_949</td>\n",
" <td>['procedure_code_45976', 'procedure_code_26419...</td>\n",
" <td>['medication_code_9684', 'medication_code_8486...</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>7772</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7217</th>\n",
" <td>udacity_health_encounter_id_995</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" <td>dx_code_4976</td>\n",
" <td>['procedure_code_8156', 'procedure_code_23220']</td>\n",
" <td>['medication_code_5388', 'medication_code_5533...</td>\n",
" <td>['lab_code_2049']</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>995</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID PATIENT_ID \\\n",
"94 udacity_health_encounter_id_1091 udacity_health_patient_id_309 \n",
"450 udacity_health_encounter_id_1436 udacity_health_patient_id_309 \n",
"605 udacity_health_encounter_id_1582 udacity_health_patient_id_309 \n",
"665 udacity_health_encounter_id_1643 udacity_health_patient_id_309 \n",
"684 udacity_health_encounter_id_1662 udacity_health_patient_id_309 \n",
"867 udacity_health_encounter_id_1837 udacity_health_patient_id_309 \n",
"1024 udacity_health_encounter_id_1985 udacity_health_patient_id_309 \n",
"1042 udacity_health_encounter_id_2000 udacity_health_patient_id_309 \n",
"1196 udacity_health_encounter_id_2142 udacity_health_patient_id_309 \n",
"1289 udacity_health_encounter_id_2229 udacity_health_patient_id_309 \n",
"1340 udacity_health_encounter_id_2281 udacity_health_patient_id_309 \n",
"1596 udacity_health_encounter_id_2531 udacity_health_patient_id_309 \n",
"1702 udacity_health_encounter_id_2631 udacity_health_patient_id_309 \n",
"2069 udacity_health_encounter_id_2989 udacity_health_patient_id_309 \n",
"2291 udacity_health_encounter_id_3209 udacity_health_patient_id_309 \n",
"2507 udacity_health_encounter_id_3416 udacity_health_patient_id_309 \n",
"2696 udacity_health_encounter_id_360 udacity_health_patient_id_309 \n",
"2890 udacity_health_encounter_id_3792 udacity_health_patient_id_309 \n",
"2960 udacity_health_encounter_id_3859 udacity_health_patient_id_309 \n",
"3064 udacity_health_encounter_id_3961 udacity_health_patient_id_309 \n",
"3167 udacity_health_encounter_id_4058 udacity_health_patient_id_309 \n",
"3280 udacity_health_encounter_id_4168 udacity_health_patient_id_309 \n",
"3353 udacity_health_encounter_id_4239 udacity_health_patient_id_309 \n",
"3561 udacity_health_encounter_id_4439 udacity_health_patient_id_309 \n",
"3585 udacity_health_encounter_id_4462 udacity_health_patient_id_309 \n",
"3718 udacity_health_encounter_id_4589 udacity_health_patient_id_309 \n",
"3752 udacity_health_encounter_id_4619 udacity_health_patient_id_309 \n",
"3916 udacity_health_encounter_id_478 udacity_health_patient_id_309 \n",
"4552 udacity_health_encounter_id_5398 udacity_health_patient_id_309 \n",
"4687 udacity_health_encounter_id_553 udacity_health_patient_id_309 \n",
"4771 udacity_health_encounter_id_5611 udacity_health_patient_id_309 \n",
"4785 udacity_health_encounter_id_5624 udacity_health_patient_id_309 \n",
"4796 udacity_health_encounter_id_5635 udacity_health_patient_id_309 \n",
"4862 udacity_health_encounter_id_5701 udacity_health_patient_id_309 \n",
"4881 udacity_health_encounter_id_5721 udacity_health_patient_id_309 \n",
"5132 udacity_health_encounter_id_597 udacity_health_patient_id_309 \n",
"5306 udacity_health_encounter_id_6139 udacity_health_patient_id_309 \n",
"5385 udacity_health_encounter_id_622 udacity_health_patient_id_309 \n",
"5472 udacity_health_encounter_id_6305 udacity_health_patient_id_309 \n",
"5564 udacity_health_encounter_id_6394 udacity_health_patient_id_309 \n",
"5669 udacity_health_encounter_id_6497 udacity_health_patient_id_309 \n",
"5684 udacity_health_encounter_id_6512 udacity_health_patient_id_309 \n",
"5829 udacity_health_encounter_id_6662 udacity_health_patient_id_309 \n",
"5846 udacity_health_encounter_id_668 udacity_health_patient_id_309 \n",
"5946 udacity_health_encounter_id_6777 udacity_health_patient_id_309 \n",
"6508 udacity_health_encounter_id_7327 udacity_health_patient_id_309 \n",
"6892 udacity_health_encounter_id_7696 udacity_health_patient_id_309 \n",
"6969 udacity_health_encounter_id_7772 udacity_health_patient_id_309 \n",
"7217 udacity_health_encounter_id_995 udacity_health_patient_id_309 \n",
"\n",
" PRINCIPAL_DIAGNOSIS_CODE \\\n",
"94 dx_code_62502 \n",
"450 dx_code_37825 \n",
"605 dx_code_79663 \n",
"665 dx_code_60773 \n",
"684 dx_code_29272 \n",
"867 dx_code_56368 \n",
"1024 dx_code_38581 \n",
"1042 dx_code_91230 \n",
"1196 dx_code_58527 \n",
"1289 dx_code_46826 \n",
"1340 dx_code_39484 \n",
"1596 dx_code_75423 \n",
"1702 dx_code_38210 \n",
"2069 dx_code_80145 \n",
"2291 dx_code_74710 \n",
"2507 dx_code_43343 \n",
"2696 dx_code_94868 \n",
"2890 dx_code_37738 \n",
"2960 dx_code_17144 \n",
"3064 dx_code_61465 \n",
"3167 dx_code_73636 \n",
"3280 dx_code_9795 \n",
"3353 dx_code_28764 \n",
"3561 dx_code_31393 \n",
"3585 dx_code_53624 \n",
"3718 dx_code_67438 \n",
"3752 dx_code_79460 \n",
"3916 dx_code_43778 \n",
"4552 dx_code_58513 \n",
"4687 dx_code_88408 \n",
"4771 dx_code_46410 \n",
"4785 dx_code_96267 \n",
"4796 dx_code_67422 \n",
"4862 dx_code_4007 \n",
"4881 dx_code_11147 \n",
"5132 dx_code_25907 \n",
"5306 dx_code_40784 \n",
"5385 dx_code_25556 \n",
"5472 dx_code_93595 \n",
"5564 dx_code_17875 \n",
"5669 dx_code_36095 \n",
"5684 dx_code_92784 \n",
"5829 dx_code_26279 \n",
"5846 dx_code_22666 \n",
"5946 dx_code_78269 \n",
"6508 dx_code_18813 \n",
"6892 dx_code_24785 \n",
"6969 dx_code_949 \n",
"7217 dx_code_4976 \n",
"\n",
" PROCEDURE_CODE \\\n",
"94 ['procedure_code_69597'] \n",
"450 ['procedure_code_14044', 'procedure_code_47911... \n",
"605 [] \n",
"665 ['procedure_code_39999', 'procedure_code_31842... \n",
"684 ['procedure_code_12589', 'procedure_code_3005'... \n",
"867 ['procedure_code_14069', 'procedure_code_15230... \n",
"1024 ['procedure_code_1431', 'procedure_code_71061'] \n",
"1042 [] \n",
"1196 ['procedure_code_56979', 'procedure_code_24797... \n",
"1289 ['procedure_code_50916', 'procedure_code_20032... \n",
"1340 ['procedure_code_33400', 'procedure_code_56', ... \n",
"1596 ['procedure_code_15600', 'procedure_code_19008... \n",
"1702 ['procedure_code_121'] \n",
"2069 ['procedure_code_28893', 'procedure_code_9124'... \n",
"2291 ['procedure_code_41970', 'procedure_code_64032... \n",
"2507 ['procedure_code_11834', 'procedure_code_12304... \n",
"2696 [] \n",
"2890 ['procedure_code_11852', 'procedure_code_45278... \n",
"2960 ['procedure_code_69917'] \n",
"3064 ['procedure_code_15248', 'procedure_code_65724... \n",
"3167 ['procedure_code_5646', 'procedure_code_1381',... \n",
"3280 ['procedure_code_44804', 'procedure_code_10917... \n",
"3353 ['procedure_code_31873', 'procedure_code_29930... \n",
"3561 ['procedure_code_69203', 'procedure_code_53820... \n",
"3585 ['procedure_code_64278', 'procedure_code_19230'] \n",
"3718 [] \n",
"3752 ['procedure_code_8648', 'procedure_code_22388'... \n",
"3916 [] \n",
"4552 ['procedure_code_3629', 'procedure_code_56509'... \n",
"4687 ['procedure_code_19471', 'procedure_code_17060... \n",
"4771 ['procedure_code_71656', 'procedure_code_29912... \n",
"4785 [] \n",
"4796 [] \n",
"4862 ['procedure_code_8263', 'procedure_code_7509'] \n",
"4881 ['procedure_code_68127', 'procedure_code_3768'... \n",
"5132 ['procedure_code_41138', 'procedure_code_28596'] \n",
"5306 ['procedure_code_48973', 'procedure_code_7953'... \n",
"5385 ['procedure_code_10462'] \n",
"5472 ['procedure_code_6005', 'procedure_code_13203'... \n",
"5564 ['procedure_code_42424'] \n",
"5669 ['procedure_code_33881', 'procedure_code_6396'... \n",
"5684 ['procedure_code_31090', 'procedure_code_28839... \n",
"5829 ['procedure_code_47253', 'procedure_code_52416... \n",
"5846 ['procedure_code_10507', 'procedure_code_39636... \n",
"5946 ['procedure_code_58888', 'procedure_code_72745'] \n",
"6508 ['procedure_code_9329', 'procedure_code_28542'... \n",
"6892 ['procedure_code_31137', 'procedure_code_23811... \n",
"6969 ['procedure_code_45976', 'procedure_code_26419... \n",
"7217 ['procedure_code_8156', 'procedure_code_23220'] \n",
"\n",
" MEDICATION_CODE \\\n",
"94 ['medication_code_9253', 'medication_code_7444... \n",
"450 ['medication_code_4081', 'medication_code_6050... \n",
"605 ['medication_code_4450', 'medication_code_4900... \n",
"665 ['medication_code_413', 'medication_code_5966'... \n",
"684 ['medication_code_9845', 'medication_code_6602... \n",
"867 ['medication_code_5820', 'medication_code_8028... \n",
"1024 ['medication_code_3134', 'medication_code_4868... \n",
"1042 [] \n",
"1196 ['medication_code_7631', 'medication_code_7508... \n",
"1289 ['medication_code_9204', 'medication_code_1282... \n",
"1340 ['medication_code_1938', 'medication_code_4076... \n",
"1596 ['medication_code_7265', 'medication_code_2247... \n",
"1702 ['medication_code_7480', 'medication_code_7917... \n",
"2069 ['medication_code_43', 'medication_code_5602',... \n",
"2291 ['medication_code_7596', 'medication_code_7481... \n",
"2507 ['medication_code_3767', 'medication_code_7630... \n",
"2696 ['medication_code_193', 'medication_code_3556'... \n",
"2890 ['medication_code_7882', 'medication_code_6562... \n",
"2960 ['medication_code_2152', 'medication_code_7078... \n",
"3064 ['medication_code_701', 'medication_code_8324'... \n",
"3167 ['medication_code_4201', 'medication_code_5328... \n",
"3280 ['medication_code_2738', 'medication_code_9919... \n",
"3353 ['medication_code_7607', 'medication_code_6643... \n",
"3561 ['medication_code_9780', 'medication_code_7008... \n",
"3585 ['medication_code_8759', 'medication_code_59',... \n",
"3718 ['medication_code_8103'] \n",
"3752 ['medication_code_4773', 'medication_code_5839... \n",
"3916 ['medication_code_1646', 'medication_code_4346'] \n",
"4552 ['medication_code_5697', 'medication_code_9845... \n",
"4687 ['medication_code_1541', 'medication_code_5218... \n",
"4771 ['medication_code_3365', 'medication_code_862'... \n",
"4785 ['medication_code_1343'] \n",
"4796 ['medication_code_2864'] \n",
"4862 ['medication_code_3925', 'medication_code_3163'] \n",
"4881 ['medication_code_4722', 'medication_code_2681... \n",
"5132 ['medication_code_4777', 'medication_code_4318... \n",
"5306 ['medication_code_3191', 'medication_code_105'... \n",
"5385 ['medication_code_3626', 'medication_code_311'] \n",
"5472 ['medication_code_4513'] \n",
"5564 [] \n",
"5669 ['medication_code_6849', 'medication_code_2364... \n",
"5684 ['medication_code_5274', 'medication_code_6976... \n",
"5829 ['medication_code_2602', 'medication_code_3320... \n",
"5846 ['medication_code_4676', 'medication_code_4760... \n",
"5946 ['medication_code_227', 'medication_code_7799'] \n",
"6508 ['medication_code_4006', 'medication_code_485'... \n",
"6892 ['medication_code_2626', 'medication_code_6310... \n",
"6969 ['medication_code_9684', 'medication_code_8486... \n",
"7217 ['medication_code_5388', 'medication_code_5533... \n",
"\n",
" LAB_CODE \\\n",
"94 ['lab_code_9415', 'lab_code_8179'] \n",
"450 ['lab_code_6246', 'lab_code_7197'] \n",
"605 [] \n",
"665 ['lab_code_4295', 'lab_code_2383'] \n",
"684 ['lab_code_4330', 'lab_code_3499', 'lab_code_1... \n",
"867 ['lab_code_8964'] \n",
"1024 ['lab_code_7702'] \n",
"1042 ['lab_code_8062'] \n",
"1196 ['lab_code_3908', 'lab_code_1500'] \n",
"1289 ['lab_code_5607', 'lab_code_5139'] \n",
"1340 ['lab_code_9178', 'lab_code_3557', 'lab_code_2... \n",
"1596 ['lab_code_8336', 'lab_code_7292', 'lab_code_8... \n",
"1702 ['lab_code_4777'] \n",
"2069 ['lab_code_3392', 'lab_code_9665', 'lab_code_4... \n",
"2291 ['lab_code_9105'] \n",
"2507 ['lab_code_774', 'lab_code_9591', 'lab_code_31... \n",
"2696 [] \n",
"2890 ['lab_code_3787', 'lab_code_6452', 'lab_code_6... \n",
"2960 ['lab_code_5682'] \n",
"3064 ['lab_code_7691', 'lab_code_4598', 'lab_code_1... \n",
"3167 ['lab_code_9083', 'lab_code_1392', 'lab_code_1... \n",
"3280 ['lab_code_4690', 'lab_code_5570', 'lab_code_2... \n",
"3353 ['lab_code_8492', 'lab_code_3309', 'lab_code_6... \n",
"3561 [] \n",
"3585 ['lab_code_3790', 'lab_code_591', 'lab_code_89... \n",
"3718 [] \n",
"3752 ['lab_code_5055'] \n",
"3916 [] \n",
"4552 ['lab_code_4909'] \n",
"4687 ['lab_code_4271', 'lab_code_7946', 'lab_code_9... \n",
"4771 ['lab_code_8280', 'lab_code_9803', 'lab_code_1... \n",
"4785 ['lab_code_9589'] \n",
"4796 [] \n",
"4862 [] \n",
"4881 ['lab_code_8853', 'lab_code_3111', 'lab_code_2... \n",
"5132 ['lab_code_1043'] \n",
"5306 ['lab_code_5078', 'lab_code_2361', 'lab_code_9... \n",
"5385 ['lab_code_4960'] \n",
"5472 ['lab_code_7349', 'lab_code_1043', 'lab_code_9... \n",
"5564 ['lab_code_5222'] \n",
"5669 ['lab_code_2570', 'lab_code_2383', 'lab_code_8... \n",
"5684 ['lab_code_4639'] \n",
"5829 ['lab_code_5441'] \n",
"5846 ['lab_code_7518', 'lab_code_1774', 'lab_code_5... \n",
"5946 [] \n",
"6508 ['lab_code_71', 'lab_code_7769', 'lab_code_659... \n",
"6892 ['lab_code_8317', 'lab_code_2116', 'lab_code_2... \n",
"6969 [] \n",
"7217 ['lab_code_2049'] \n",
"\n",
" LABEL ENCOUNTER_ID_NUMBER \n",
"94 [0, 0, 0, 0, 0, 0, 0] 1091 \n",
"450 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 1436 \n",
"605 [0, 0, 0] 1582 \n",
"665 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 1643 \n",
"684 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 1662 \n",
"867 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 1837 \n",
"1024 [0, 0, 0, 0, 0, 0, 0, 0, 0] 1985 \n",
"1042 [0] 2000 \n",
"1196 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 2142 \n",
"1289 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 2229 \n",
"1340 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 2281 \n",
"1596 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 2531 \n",
"1702 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 2631 \n",
"2069 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 2989 \n",
"2291 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 3209 \n",
"2507 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 3416 \n",
"2696 [0, 0, 0] 360 \n",
"2890 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 3792 \n",
"2960 [0, 0, 0, 0, 0, 0, 0] 3859 \n",
"3064 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 3961 \n",
"3167 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 4058 \n",
"3280 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 4168 \n",
"3353 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 4239 \n",
"3561 [0, 0, 0, 0, 0, 0, 0, 0, 0] 4439 \n",
"3585 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 4462 \n",
"3718 [0] 4589 \n",
"3752 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 4619 \n",
"3916 [0, 0] 478 \n",
"4552 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 5398 \n",
"4687 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 553 \n",
"4771 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 5611 \n",
"4785 [0, 0] 5624 \n",
"4796 [0] 5635 \n",
"4862 [0, 0, 0, 0] 5701 \n",
"4881 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 5721 \n",
"5132 [0, 0, 0, 0, 0, 0] 597 \n",
"5306 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 6139 \n",
"5385 [0, 0, 0, 0] 622 \n",
"5472 [0, 0, 0, 0, 0, 0, 0] 6305 \n",
"5564 [0, 0] 6394 \n",
"5669 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 6497 \n",
"5684 [0, 0, 0, 0, 0, 0, 0, 0, 0] 6512 \n",
"5829 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 6662 \n",
"5846 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 668 \n",
"5946 [0, 0, 0, 0] 6777 \n",
"6508 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 7327 \n",
"6892 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 7696 \n",
"6969 [0, 0, 0, 0, 0, 0, 0] 7772 \n",
"7217 [0, 0, 0, 0, 0, 0, 0] 995 "
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"ehr_level_df[ehr_level_df['PATIENT_ID']=='udacity_health_patient_id_309']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For patient id 309, the selected encounter should be 7772."
]
},
{
"cell_type": "code",
"execution_count": 10,
"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>ENCOUNTER_ID</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" <th>PATIENT_ID</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>6969</th>\n",
" <td>udacity_health_encounter_id_7772</td>\n",
" <td>7772</td>\n",
" <td>udacity_health_patient_id_309</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID ENCOUNTER_ID_NUMBER \\\n",
"6969 udacity_health_encounter_id_7772 7772 \n",
"\n",
" PATIENT_ID \n",
"6969 udacity_health_patient_id_309 "
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_last_encounter_df[test_last_encounter_df['PATIENT_ID']=='udacity_health_patient_id_309']"
]
},
{
"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>ENCOUNTER_ID</th>\n",
" <th>PATIENT_ID</th>\n",
" <th>PRINCIPAL_DIAGNOSIS_CODE</th>\n",
" <th>PROCEDURE_CODE</th>\n",
" <th>MEDICATION_CODE</th>\n",
" <th>LAB_CODE</th>\n",
" <th>LABEL</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>263</th>\n",
" <td>udacity_health_encounter_id_1258</td>\n",
" <td>udacity_health_patient_id_418</td>\n",
" <td>dx_code_87320</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_7475']</td>\n",
" <td>[]</td>\n",
" <td>[0]</td>\n",
" <td>1258</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1230</th>\n",
" <td>udacity_health_encounter_id_2173</td>\n",
" <td>udacity_health_patient_id_418</td>\n",
" <td>dx_code_72104</td>\n",
" <td>['procedure_code_37393', 'procedure_code_48757...</td>\n",
" <td>['medication_code_3809', 'medication_code_9141...</td>\n",
" <td>['lab_code_2839', 'lab_code_2665', 'lab_code_3...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
" <td>2173</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2448</th>\n",
" <td>udacity_health_encounter_id_3362</td>\n",
" <td>udacity_health_patient_id_418</td>\n",
" <td>dx_code_77979</td>\n",
" <td>['procedure_code_64397']</td>\n",
" <td>['medication_code_2807', 'medication_code_6979']</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0]</td>\n",
" <td>3362</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID PATIENT_ID \\\n",
"263 udacity_health_encounter_id_1258 udacity_health_patient_id_418 \n",
"1230 udacity_health_encounter_id_2173 udacity_health_patient_id_418 \n",
"2448 udacity_health_encounter_id_3362 udacity_health_patient_id_418 \n",
"\n",
" PRINCIPAL_DIAGNOSIS_CODE \\\n",
"263 dx_code_87320 \n",
"1230 dx_code_72104 \n",
"2448 dx_code_77979 \n",
"\n",
" PROCEDURE_CODE \\\n",
"263 [] \n",
"1230 ['procedure_code_37393', 'procedure_code_48757... \n",
"2448 ['procedure_code_64397'] \n",
"\n",
" MEDICATION_CODE \\\n",
"263 ['medication_code_7475'] \n",
"1230 ['medication_code_3809', 'medication_code_9141... \n",
"2448 ['medication_code_2807', 'medication_code_6979'] \n",
"\n",
" LAB_CODE \\\n",
"263 [] \n",
"1230 ['lab_code_2839', 'lab_code_2665', 'lab_code_3... \n",
"2448 [] \n",
"\n",
" LABEL ENCOUNTER_ID_NUMBER \n",
"263 [0] 1258 \n",
"1230 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ... 2173 \n",
"2448 [0, 0, 0] 3362 "
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"ehr_level_df[ehr_level_df['PATIENT_ID']=='udacity_health_patient_id_418']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For patient id 418, the selected encounter should be 3362."
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>ENCOUNTER_ID</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" <th>PATIENT_ID</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2448</th>\n",
" <td>udacity_health_encounter_id_3362</td>\n",
" <td>3362</td>\n",
" <td>udacity_health_patient_id_418</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID ENCOUNTER_ID_NUMBER \\\n",
"2448 udacity_health_encounter_id_3362 3362 \n",
"\n",
" PATIENT_ID \n",
"2448 udacity_health_patient_id_418 "
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_last_encounter_df[test_last_encounter_df['PATIENT_ID']=='udacity_health_patient_id_418']"
]
},
{
"cell_type": "code",
"execution_count": 13,
"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>ENCOUNTER_ID</th>\n",
" <th>PATIENT_ID</th>\n",
" <th>PRINCIPAL_DIAGNOSIS_CODE</th>\n",
" <th>PROCEDURE_CODE</th>\n",
" <th>MEDICATION_CODE</th>\n",
" <th>LAB_CODE</th>\n",
" <th>LABEL</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>1678</th>\n",
" <td>udacity_health_encounter_id_2608</td>\n",
" <td>udacity_health_patient_id_908</td>\n",
" <td>dx_code_52877</td>\n",
" <td>[]</td>\n",
" <td>['medication_code_6189']</td>\n",
" <td>[]</td>\n",
" <td>[0]</td>\n",
" <td>2608</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2796</th>\n",
" <td>udacity_health_encounter_id_3698</td>\n",
" <td>udacity_health_patient_id_908</td>\n",
" <td>dx_code_52688</td>\n",
" <td>['procedure_code_60912']</td>\n",
" <td>[]</td>\n",
" <td>['lab_code_4691', 'lab_code_9380']</td>\n",
" <td>[0, 0, 0]</td>\n",
" <td>3698</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5301</th>\n",
" <td>udacity_health_encounter_id_6132</td>\n",
" <td>udacity_health_patient_id_908</td>\n",
" <td>dx_code_36066</td>\n",
" <td>['procedure_code_5724', 'procedure_code_55191'...</td>\n",
" <td>['medication_code_9532', 'medication_code_3398...</td>\n",
" <td>['lab_code_9732', 'lab_code_2808', 'lab_code_3...</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>6132</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7131</th>\n",
" <td>udacity_health_encounter_id_910</td>\n",
" <td>udacity_health_patient_id_908</td>\n",
" <td>dx_code_98539</td>\n",
" <td>['procedure_code_8188', 'procedure_code_59872'...</td>\n",
" <td>['medication_code_5112', 'medication_code_3537...</td>\n",
" <td>[]</td>\n",
" <td>[0, 0, 0, 0, 0, 0, 0]</td>\n",
" <td>910</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID PATIENT_ID \\\n",
"1678 udacity_health_encounter_id_2608 udacity_health_patient_id_908 \n",
"2796 udacity_health_encounter_id_3698 udacity_health_patient_id_908 \n",
"5301 udacity_health_encounter_id_6132 udacity_health_patient_id_908 \n",
"7131 udacity_health_encounter_id_910 udacity_health_patient_id_908 \n",
"\n",
" PRINCIPAL_DIAGNOSIS_CODE \\\n",
"1678 dx_code_52877 \n",
"2796 dx_code_52688 \n",
"5301 dx_code_36066 \n",
"7131 dx_code_98539 \n",
"\n",
" PROCEDURE_CODE \\\n",
"1678 [] \n",
"2796 ['procedure_code_60912'] \n",
"5301 ['procedure_code_5724', 'procedure_code_55191'... \n",
"7131 ['procedure_code_8188', 'procedure_code_59872'... \n",
"\n",
" MEDICATION_CODE \\\n",
"1678 ['medication_code_6189'] \n",
"2796 [] \n",
"5301 ['medication_code_9532', 'medication_code_3398... \n",
"7131 ['medication_code_5112', 'medication_code_3537... \n",
"\n",
" LAB_CODE \\\n",
"1678 [] \n",
"2796 ['lab_code_4691', 'lab_code_9380'] \n",
"5301 ['lab_code_9732', 'lab_code_2808', 'lab_code_3... \n",
"7131 [] \n",
"\n",
" LABEL ENCOUNTER_ID_NUMBER \n",
"1678 [0] 2608 \n",
"2796 [0, 0, 0] 3698 \n",
"5301 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] 6132 \n",
"7131 [0, 0, 0, 0, 0, 0, 0] 910 "
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"ehr_level_df[ehr_level_df['PATIENT_ID']=='udacity_health_patient_id_908']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For patient id 908, the selected encounter should be 6132."
]
},
{
"cell_type": "code",
"execution_count": 14,
"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>ENCOUNTER_ID</th>\n",
" <th>ENCOUNTER_ID_NUMBER</th>\n",
" <th>PATIENT_ID</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>5301</th>\n",
" <td>udacity_health_encounter_id_6132</td>\n",
" <td>6132</td>\n",
" <td>udacity_health_patient_id_908</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" ENCOUNTER_ID ENCOUNTER_ID_NUMBER \\\n",
"5301 udacity_health_encounter_id_6132 6132 \n",
"\n",
" PATIENT_ID \n",
"5301 udacity_health_patient_id_908 "
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_last_encounter_df[test_last_encounter_df['PATIENT_ID']=='udacity_health_patient_id_908']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Exercise L3 - 2: Dataset Splitting"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Instructions\n",
"- Split the provided dataset into a train and test split but be sure not to mix patient encounter records across the two partitions\n",
"- Be sure to run the following three tests\n",
" - Patient data in only one partition\n",
" - Total unique number of patients across all partitions = total number unique patients in original full dataset\n",
" - Total number of rows original dataset = sum of rows across splits"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [],
"source": [
"splitting_exercise_dataset_path = \"./data/SYNTHETIC_EHR_DATASET.csv\""
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Solution"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This is largely a review of two parts in this lesson and you can use most of the same code for each step. The key is to identify the level of the dataset and then to convert it to the encounter level before you do your splits. Then perform the splitting and run the tests."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Convert to Encounter Level"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [],
"source": [
"# convert to encounter and then split but make sure \n",
"ehr_pre_split_df = pd.read_csv(splitting_exercise_dataset_path)\n",
"grouping_field_list = ['ENCOUNTER_ID', 'PATIENT_ID', 'PRINCIPAL_DIAGNOSIS_CODE']\n",
"non_grouped_field_list = [c for c in ehr_pre_split_df.columns if c not in grouping_field_list]\n",
"ehr_encounter_df = ehr_pre_split_df.groupby(grouping_field_list)[non_grouped_field_list].agg(lambda x: \n",
" list([y for y in x if y is not np.nan ] ) ).reset_index()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Split at Patient Level"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [],
"source": [
"PATIENT_ID_FIELD = 'PATIENT_ID'\n",
"TEST_PERCENTAGE = 0.2"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [],
"source": [
"def split_dataset_patient_level(df, key, test_percentage=0.2):\n",
" df = df.iloc[np.random.permutation(len(df))]\n",
" unique_values = df[key].unique()\n",
" total_values = len(unique_values)\n",
" sample_size = round(total_values * (1 - test_percentage ))\n",
" train = df[df[key].isin(unique_values[:sample_size])].reset_index(drop=True)\n",
" test = df[df[key].isin(unique_values[sample_size:])].reset_index(drop=True)\n",
" return train, test"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"train_df, test_df = split_dataset_patient_level(ehr_encounter_df, PATIENT_ID_FIELD, TEST_PERCENTAGE)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Test passed for patient data in only one partition\n"
]
}
],
"source": [
"assert len(set(train_df[PATIENT_ID_FIELD].unique()).intersection(set(test_df[PATIENT_ID_FIELD].unique()))) == 0\n",
"print(\"Test passed for patient data in only one partition\")"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Test passed for number of unique patients being equal!\n"
]
}
],
"source": [
"assert (train_df[PATIENT_ID_FIELD].nunique() + test_df[PATIENT_ID_FIELD].nunique()) == ehr_encounter_df[PATIENT_ID_FIELD].nunique()\n",
"print(\"Test passed for number of unique patients being equal!\")"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Test passed for number of total rows equal!\n"
]
}
],
"source": [
"assert len(train_df) + len(test_df) == len(ehr_encounter_df)\n",
"print(\"Test passed for number of total rows equal!\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Optional\n",
"- Check label distribution and use scikitlearn - https://scikit-learn.org/stable/auto_examples/model_selection/plot_cv_indices.html#sphx-glr-auto-examples-model-selection-plot-cv-indices-py"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Exercise L3 - 3: Build Bucketed Numeric Feature with TF"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Instructions\n",
"- Given the Swiss heart disease dataset that we worked with earlier, build a bucketed numeric feature from the age feature. \n",
"- For this exercise, use the Tensorflow csv function for loading the dataset directly into a TF tensor -https://www.tensorflow.org/api_docs/python/tf/data/experimental/make_csv_dataset. This approach will be useful for when you have much larger datasets and also allows you to bypass loading the dataset in Pandas.\n",
"- More information on the Tensorflow bucketized feature can be found here https://www.tensorflow.org/api_docs/python/tf/feature_column/bucketized_column. Bucketed features take as input the numeric feature that we covered in the lesson. For the numeric feature, you do not need to normalize it like we did in the lesson. "
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"import tensorflow as tf\n",
"swiss_dataset_path = \"./data/lesson_exercise_swiss_dataset.csv\"\n",
"BATCH_SIZE =128\n",
"PREDICTOR_FIELD = 'num_label'"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Solution"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [],
"source": [
"# ETL with TF dataset make csv function\n",
"swiss_tf_dataset = tf.data.experimental.make_csv_dataset( swiss_dataset_path, batch_size=BATCH_SIZE, \n",
" num_epochs=1, label_name=PREDICTOR_FIELD, header=True)"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"OrderedDict([('age',\n",
" <tf.Tensor: shape=(123,), dtype=int32, numpy=\n",
" array([65, 70, 38, 52, 51, 60, 56, 53, 51, 56, 61, 59, 45, 62, 50, 54, 61,\n",
" 38, 61, 60, 53, 64, 65, 38, 58, 66, 62, 58, 68, 58, 52, 55, 32, 61,\n",
" 48, 54, 50, 43, 50, 51, 47, 65, 42, 53, 67, 55, 43, 61, 73, 54, 69,\n",
" 46, 64, 65, 60, 35, 52, 63, 56, 38, 61, 59, 61, 57, 57, 66, 62, 51,\n",
" 42, 36, 59, 57, 72, 64, 53, 60, 56, 47, 60, 63, 47, 53, 63, 63, 56,\n",
" 62, 56, 40, 59, 53, 70, 50, 51, 43, 41, 62, 34, 68, 62, 61, 60, 55,\n",
" 74, 69, 55, 56, 57, 52, 38, 53, 38, 62, 53, 61, 56, 57, 51, 46, 59,\n",
" 63, 51, 57, 47], dtype=int32)>)])"
]
},
"execution_count": 25,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"swiss_dataset_batch = next(iter(swiss_tf_dataset))[0]\n",
"swiss_dataset_batch"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [],
"source": [
"# create TF numeric feature\n",
"tf_numeric_age_feature = tf.feature_column.numeric_column(key='age', default_value=0, dtype=tf.float64)"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [],
"source": [
"#boundaries for the different age buckets\n",
"b_list = [ 0, 18, 25, 40, 55, 65, 80, 100]\n",
"#create TF bucket feature from numeric feature\n",
"tf_bucket_age_feature = tf.feature_column.bucketized_column(source_column=tf_numeric_age_feature, boundaries= b_list)"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [],
"source": [
"def demo(feature_column, example_batch):\n",
" feature_layer = tf.keras.layers.DenseFeatures(feature_column)\n",
" print(feature_layer(example_batch))\n",
" print(\"\\nExample of one transformed row:\")\n",
" print(feature_layer(example_batch).numpy()[0])"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Example bucket field:\n",
"BucketizedColumn(source_column=NumericColumn(key='age', shape=(1,), default_value=(0,), dtype=tf.float64, normalizer_fn=None), boundaries=(0, 18, 25, 40, 55, 65, 80, 100))\n",
"\n",
"tf.Tensor(\n",
"[[0. 0. 0. ... 1. 0. 0.]\n",
" [0. 0. 0. ... 1. 0. 0.]\n",
" [0. 0. 0. ... 0. 0. 0.]\n",
" ...\n",
" [0. 0. 0. ... 0. 0. 0.]\n",
" [0. 0. 0. ... 0. 0. 0.]\n",
" [0. 0. 0. ... 0. 0. 0.]], shape=(123, 9), dtype=float32)\n",
"\n",
"Example of one transformed row:\n",
"[0. 0. 0. 0. 0. 0. 1. 0. 0.]\n"
]
}
],
"source": [
"print(\"Example bucket field:\\n{}\\n\".format(tf_bucket_age_feature))\n",
"demo(tf_bucket_age_feature, swiss_dataset_batch)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Exercise L3 - 4: Build Embedding Categorical Feature with TF"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Instructions\n",
"- Build a 10 dimension embedding feature for the PRINCIPAL_DIAGNOSIS_CODE field\n",
"- Here is the link to the Tensorflow Embedding column documentation -https://www.tensorflow.org/api_docs/python/tf/feature_column/embedding_column\n",
"- Some functions provided below to assist"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [],
"source": [
"ehr_line_df = pd.read_csv(\"./data/SYNTHETIC_EHR_DATASET.csv\")\n",
"cat_example_df = ehr_line_df[['ENCOUNTER_ID', 'PRINCIPAL_DIAGNOSIS_CODE', 'LABEL']]"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [],
"source": [
"#adapted from https://www.tensorflow.org/tutorials/structured_data/feature_columns\n",
"def df_to_dataset(df, predictor, batch_size=32):\n",
" df = df.copy()\n",
" labels = df.pop(predictor)\n",
" ds = tf.data.Dataset.from_tensor_slices((dict(df), labels))\n",
" ds = ds.shuffle(buffer_size=len(df))\n",
" ds = ds.batch(batch_size)\n",
" return ds"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [],
"source": [
"BATCH_SIZE = 64\n",
"PREDICTOR_FIELD = 'LABEL'\n",
"categorical_tf_ds = df_to_dataset(cat_example_df, PREDICTOR_FIELD, batch_size=BATCH_SIZE)"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [],
"source": [
"# build vocab for categorical features\n",
"def write_vocabulary_file(vocab_list, field_name, default_value, vocab_dir='./vocab/'):\n",
" output_file_path = os.path.join(vocab_dir, str(field_name) + \"_vocab.txt\")\n",
" # put default value in first row as TF requires\n",
" vocab_list = np.insert(vocab_list, 0, default_value, axis=0) \n",
" df = pd.DataFrame(vocab_list).to_csv(output_file_path, index=None, header=None)\n",
" return output_file_path\n",
"\n",
"def build_vocab_files(df, categorical_column_list, default_value='00'):\n",
" vocab_files_list = []\n",
" for c in categorical_column_list:\n",
" v_file = write_vocabulary_file(df[c].unique(), c, default_value)\n",
" vocab_files_list.append(v_file)\n",
" return vocab_files_list"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Solution"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"# add logic to add if not exist\n",
"#os.mkdir(\"./vocab/\")"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'./vocab/PRINCIPAL_DIAGNOSIS_CODE_vocab.txt'"
]
},
"execution_count": 35,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"categorical_field_list = [\"PRINCIPAL_DIAGNOSIS_CODE\"]\n",
"vocab_files_list = build_vocab_files(cat_example_df, categorical_field_list)\n",
"vocab_files_list[0]"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFO:tensorflow:vocabulary_size = 6753 in PRINCIPAL_DIAGNOSIS_CODE is inferred from the number of elements in the vocabulary_file ./vocab/PRINCIPAL_DIAGNOSIS_CODE_vocab.txt.\n"
]
}
],
"source": [
"principal_diagnosis_vocab = tf.feature_column.categorical_column_with_vocabulary_file(\n",
" key=\"PRINCIPAL_DIAGNOSIS_CODE\", vocabulary_file = vocab_files_list[0], num_oov_buckets=1)"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [],
"source": [
"dims = 10\n",
"cat_embedded = tf.feature_column.embedding_column(principal_diagnosis_vocab, dimension=dims)"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [],
"source": [
"categorical_tf_ds_batch = next(iter(categorical_tf_ds))[0]"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"tf.Tensor(\n",
"[[ 2.21173733e-01 -1.72159821e-02 1.95488617e-01 3.94141227e-01\n",
" -1.67326014e-02 -2.14374363e-01 1.35068595e-01 -2.46864319e-01\n",
" -2.92306662e-01 1.10351130e-01]\n",
" [-3.70135516e-01 2.29885086e-01 5.77204525e-01 -3.33219431e-02\n",
" -5.73716164e-01 -2.24834934e-01 -5.42964377e-02 2.17185676e-01\n",
" 1.66468963e-01 -2.94320621e-02]\n",
" [ 2.23573312e-01 -5.65836057e-02 -8.03322047e-02 9.44105014e-02\n",
" -4.51478630e-01 -1.10069960e-01 5.88697612e-01 1.56490564e-01\n",
" -4.58488345e-01 5.05169749e-01]\n",
" [ 1.90142587e-01 6.30900711e-02 1.27180293e-01 -2.40399882e-01\n",
" -3.38476181e-01 1.58759262e-02 -1.73723385e-01 1.54326290e-01\n",
" -1.36973143e-01 -8.69284198e-02]\n",
" [ 1.09519929e-01 2.49608710e-01 -2.60479212e-01 1.12497471e-01\n",
" 1.42827034e-01 4.49611813e-01 8.05474147e-02 5.62058032e-01\n",
" 4.47835207e-01 -5.23728132e-02]\n",
" [-3.56267035e-01 -2.91776627e-01 4.70154822e-01 1.82453468e-01\n",
" -2.62245953e-01 -3.69706720e-01 -2.10007370e-01 2.95164227e-01\n",
" -1.79117620e-01 1.87984556e-02]\n",
" [ 5.47165930e-01 -3.71892065e-01 5.82392216e-01 8.78551975e-02\n",
" -2.24966273e-01 -2.86136597e-01 7.21387519e-03 -5.66901900e-02\n",
" 3.21903884e-01 -2.59898696e-02]\n",
" [ 1.45990640e-01 -2.62489647e-01 -9.33190808e-02 1.46773815e-01\n",
" -3.21463913e-01 6.96207583e-02 4.30907428e-01 -1.83946386e-01\n",
" 4.00100976e-01 1.49181694e-01]\n",
" [ 1.65900216e-01 6.07954226e-02 4.39754426e-01 3.47980261e-01\n",
" -6.31673709e-02 2.34097913e-01 6.23354614e-01 9.89933461e-02\n",
" -3.81052673e-01 2.55292118e-01]\n",
" [-3.63221377e-01 6.52561039e-02 -3.43514942e-02 4.38841879e-01\n",
" 3.45863223e-01 -5.30374110e-01 2.96102375e-01 -3.76442969e-01\n",
" -4.73127931e-01 2.56435245e-01]\n",
" [ 3.05163682e-01 1.24479391e-01 2.79286265e-01 -3.03077459e-01\n",
" 1.69501066e-01 -9.67929587e-02 -5.79946399e-01 3.07954252e-01\n",
" 1.04335368e-01 4.40943211e-01]\n",
" [ 7.73213804e-03 -5.89087903e-01 -3.68716210e-01 -1.59685805e-01\n",
" 3.60390127e-01 3.90282094e-01 5.56753218e-01 -5.86735643e-02\n",
" 2.07081214e-01 -2.53996730e-01]\n",
" [-2.05367729e-01 3.40942740e-01 2.31850058e-01 -2.97459364e-01\n",
" -3.11965346e-01 -2.48455569e-01 2.55972236e-01 -2.73559153e-01\n",
" -2.78003991e-01 -6.05599105e-01]\n",
" [ 3.62973027e-02 -5.52548230e-01 1.54541776e-01 -4.83170375e-02\n",
" 1.82704136e-01 -4.91367131e-02 3.41807932e-01 -3.83855760e-01\n",
" -3.52012664e-01 -2.12510675e-02]\n",
" [-1.12446137e-01 -4.90654647e-01 -2.00538069e-01 5.31839609e-01\n",
" -3.24222744e-02 -2.23938525e-01 2.06225201e-01 3.53894889e-01\n",
" -1.47967964e-01 -2.02423260e-01]\n",
" [ 3.21457535e-01 5.93174957e-02 -3.69869232e-01 -4.09282655e-01\n",
" 9.14927498e-02 3.32246929e-01 -2.48576999e-01 4.00003642e-01\n",
" 5.06345451e-01 2.01818004e-01]\n",
" [-1.47743538e-01 3.67682010e-01 -8.93095881e-02 9.23893377e-02\n",
" 3.89735103e-01 -3.53251211e-02 -2.62227654e-01 6.44264668e-02\n",
" -2.38305166e-01 -1.52197331e-01]\n",
" [-4.72036660e-01 -3.21255885e-02 7.68640414e-02 3.21973473e-01\n",
" 2.66465507e-02 -2.05376044e-01 3.84855568e-02 3.20453160e-02\n",
" -2.44045362e-01 -4.75476623e-01]\n",
" [-3.80438238e-01 -4.09091324e-01 2.10142478e-01 4.05305773e-01\n",
" 6.14984371e-02 3.84018242e-01 2.38912642e-01 -2.40870863e-01\n",
" -5.01134217e-01 -6.38686344e-02]\n",
" [-2.01972052e-01 5.70804775e-02 -6.07003570e-02 -1.96827278e-01\n",
" -5.08159399e-01 -1.03628360e-01 1.59624040e-01 -2.37674490e-01\n",
" 2.31359191e-02 -2.06301585e-01]\n",
" [-6.49687648e-02 -5.00375219e-02 -3.93785596e-01 -1.91538874e-02\n",
" 3.34195584e-01 -3.29131246e-01 -3.68286878e-01 -2.70653337e-01\n",
" 4.14935440e-01 2.34607071e-01]\n",
" [-5.15714847e-02 -3.64539772e-01 1.95512101e-02 1.44576281e-01\n",
" 3.24142128e-01 3.27825323e-02 3.87537144e-02 -1.59311146e-01\n",
" 1.18068174e-01 -2.08626002e-01]\n",
" [ 2.09214464e-01 4.56928276e-03 -1.04735773e-02 1.77772090e-01\n",
" -2.06444017e-03 5.46324849e-01 1.41713530e-01 4.14380163e-01\n",
" -7.64702037e-02 3.37981820e-01]\n",
" [-1.18957266e-01 4.86897141e-01 3.07010785e-02 9.68524665e-02\n",
" 6.66879192e-02 1.35202482e-01 -1.46177948e-01 -2.02172980e-01\n",
" -5.08908443e-02 1.97674543e-01]\n",
" [-1.68040857e-01 -2.06194833e-01 -3.15807194e-01 -5.24919212e-01\n",
" 2.60535568e-01 3.90114598e-02 -4.19732146e-02 3.25531751e-01\n",
" -1.92902520e-01 1.62280828e-01]\n",
" [ 3.05758744e-01 2.48721447e-02 2.02467248e-01 1.53078869e-01\n",
" -2.89428353e-01 -5.38650958e-04 2.89290726e-01 -2.21982226e-01\n",
" -4.03302729e-01 7.56065622e-02]\n",
" [-9.32436287e-02 2.44013861e-01 1.51964977e-01 1.73874572e-01\n",
" 1.94687024e-01 2.43246436e-01 -1.89445019e-01 -1.00778326e-01\n",
" 1.07042216e-01 -1.24372415e-01]\n",
" [-1.66304529e-01 2.27702618e-01 1.03454947e-01 3.77445281e-01\n",
" -1.39045894e-01 2.07322732e-01 -6.75291345e-02 4.79865782e-02\n",
" 3.46478283e-01 1.17409937e-01]\n",
" [ 4.06199872e-01 3.49460721e-01 2.44637355e-01 2.48847499e-01\n",
" 4.78596061e-01 2.15233400e-01 -6.15224898e-01 1.12012804e-01\n",
" 6.17501549e-02 -4.65285063e-01]\n",
" [ 3.34114991e-02 4.51019287e-01 -3.53240311e-01 -2.11630970e-01\n",
" -1.54820323e-01 -6.26488253e-02 -8.36589113e-02 -7.38469809e-02\n",
" 3.11094910e-01 1.83216110e-01]\n",
" [ 1.17729269e-01 2.31808484e-01 -7.75934532e-02 2.38361761e-01\n",
" -4.84732747e-01 1.08366184e-01 -1.42730400e-01 -2.09738344e-01\n",
" 3.87201577e-01 4.63078231e-01]\n",
" [ 8.02836940e-02 1.29528269e-01 -1.31656677e-02 4.83411461e-01\n",
" -1.02488682e-01 -3.96515802e-02 8.13919865e-03 -1.78335369e-01\n",
" 5.28594971e-01 -5.06561756e-01]\n",
" [ 6.97559565e-02 1.62109986e-01 2.89449006e-01 2.36351043e-02\n",
" -2.20395207e-01 1.00556277e-01 -1.82724625e-01 2.24839896e-01\n",
" 3.75871025e-02 5.97209215e-01]\n",
" [-8.72354358e-02 3.24999690e-02 -1.40968591e-01 -2.79432982e-01\n",
" -4.57580000e-01 1.81796074e-01 -1.08092166e-01 1.91744015e-01\n",
" 7.57934079e-02 3.99347872e-01]\n",
" [ 8.95942599e-02 -4.12449062e-01 -5.73571861e-01 1.75338015e-01\n",
" 5.13237193e-02 2.88603067e-01 -9.66583472e-03 -9.90523547e-02\n",
" 2.01350898e-02 -2.24946514e-01]\n",
" [ 3.46745461e-01 2.71048993e-01 -9.16679427e-02 -7.77366161e-02\n",
" 4.89920676e-01 -4.81635541e-01 -3.17499459e-01 1.99631482e-01\n",
" 2.49167159e-01 -1.57573476e-01]\n",
" [-2.30746806e-01 4.66090620e-01 4.54338253e-01 1.15700610e-01\n",
" -9.24459398e-02 -3.96088123e-01 4.27640826e-01 5.73597923e-02\n",
" -3.38077366e-01 -4.07205671e-01]\n",
" [ 3.77220392e-01 4.85850960e-01 -1.06350690e-01 2.83066966e-02\n",
" 1.84345081e-01 8.95045027e-02 -3.64821970e-01 -4.65173692e-01\n",
" -3.24850440e-01 1.91724494e-01]\n",
" [ 1.12438530e-01 -3.71120334e-01 -7.72717074e-02 5.04763909e-02\n",
" 1.85718343e-01 9.57610086e-03 -8.91020149e-02 8.43717158e-03\n",
" -2.84862280e-01 -2.92068005e-01]\n",
" [-2.34553739e-01 -8.34727511e-02 2.61431992e-01 1.10937811e-01\n",
" 1.35743394e-01 1.09667197e-01 -2.40983710e-01 1.35432109e-01\n",
" -3.67631346e-01 2.61494190e-01]\n",
" [-2.17249095e-01 3.94547760e-01 -1.96576677e-02 5.84245920e-01\n",
" -3.83032441e-01 -1.82644352e-01 -2.23389491e-01 1.96186975e-01\n",
" 5.04969656e-01 5.06922007e-01]\n",
" [-5.28254509e-01 -6.47731647e-02 -8.88943672e-02 3.63148332e-01\n",
" 1.37630776e-01 -8.34298581e-02 4.63728979e-02 -5.26239574e-02\n",
" 2.93244779e-01 4.37557340e-01]\n",
" [-8.45043510e-02 -5.66575766e-01 -1.25844344e-01 -4.33580101e-01\n",
" 4.50702727e-01 -4.36995141e-02 -4.68379319e-01 2.65026033e-01\n",
" 1.11401625e-01 2.87989408e-01]\n",
" [-1.45526037e-01 3.80581707e-01 2.30159819e-01 2.29766756e-01\n",
" -5.62798500e-01 -3.27541918e-01 5.41647911e-01 1.38492495e-01\n",
" 5.79344034e-02 -3.46691966e-01]\n",
" [ 8.84416774e-02 -5.02830684e-01 -4.80091184e-01 1.29740626e-01\n",
" -5.12592793e-01 4.52330530e-01 -7.29410127e-02 -6.56349361e-02\n",
" 2.68445611e-01 3.78060251e-01]\n",
" [ 3.10159236e-01 3.72439437e-03 1.95693165e-01 -1.10333323e-01\n",
" -1.79411083e-01 5.36282420e-01 4.06817608e-02 -3.30081224e-01\n",
" -6.11320697e-02 2.43110299e-01]\n",
" [ 6.31699383e-01 -5.95075965e-01 3.38996686e-02 -2.86349684e-01\n",
" 3.90133411e-01 -5.38905524e-02 1.67297889e-02 -4.77125466e-01\n",
" -2.29528844e-01 8.86970311e-02]\n",
" [-1.14536852e-01 -3.88067782e-01 -1.83217540e-01 -4.07678932e-01\n",
" 9.88018438e-02 3.09355021e-01 4.69066203e-01 -1.46911442e-01\n",
" -1.40690237e-01 -2.91088879e-01]\n",
" [ 3.56209844e-01 -1.63025558e-02 4.89954472e-01 -5.74329257e-01\n",
" -2.02979133e-01 4.71178532e-01 -1.29683331e-01 9.22208056e-02\n",
" -1.81245089e-01 3.09783351e-02]\n",
" [ 1.67314947e-01 2.32018098e-01 4.12859142e-01 2.12580293e-01\n",
" -1.80488937e-02 -2.63569266e-01 -1.50420174e-01 2.62736768e-01\n",
" -4.47872989e-02 7.92705640e-02]\n",
" [ 4.08414751e-01 4.20303613e-01 -4.53268826e-01 -5.07035136e-01\n",
" -4.64152284e-02 3.04141074e-01 -4.97599095e-01 5.77130355e-02\n",
" 3.46742779e-01 -2.61391848e-01]\n",
" [-1.29503995e-01 3.79304081e-01 -6.86735660e-02 -5.41104019e-01\n",
" -4.91090156e-02 -3.48790944e-01 -2.95805395e-01 8.52178186e-02\n",
" -1.33517712e-01 -7.27027562e-03]\n",
" [ 1.02716826e-01 6.80366233e-02 -4.86424804e-01 -1.81124747e-01\n",
" -3.86693597e-01 -4.50449556e-01 -2.97391951e-01 1.47938699e-01\n",
" -5.17401397e-01 8.15296099e-02]\n",
" [-5.39604016e-02 3.00839096e-01 -1.85012221e-01 3.81570131e-01\n",
" -6.21404052e-01 1.27725914e-01 2.66442865e-01 -3.82936954e-01\n",
" 2.29131177e-01 2.59222180e-01]\n",
" [ 8.62242803e-02 -5.88915162e-02 -2.91529186e-02 2.19931975e-02\n",
" 1.69219196e-01 1.95221096e-01 -3.58254582e-01 1.52518094e-01\n",
" -5.88115938e-02 2.78177381e-01]\n",
" [ 4.64907028e-02 -2.25962177e-01 -3.37895364e-01 -1.46054432e-01\n",
" -5.81172466e-01 1.34139121e-01 4.05016214e-01 -4.62647155e-02\n",
" -5.48960306e-02 -3.15349847e-01]\n",
" [ 1.98331531e-02 3.56559932e-01 3.58095616e-01 -3.89024049e-01\n",
" 3.23762357e-01 1.02973975e-01 -4.31862980e-01 1.98491976e-01\n",
" 2.03011911e-02 -8.38057250e-02]\n",
" [ 1.22427888e-01 1.81814313e-01 4.14778471e-01 -2.66290814e-01\n",
" 3.41357410e-01 -5.60003638e-01 4.35484767e-01 -3.64804953e-01\n",
" 4.21479225e-01 1.25182763e-01]\n",
" [-3.92881453e-01 3.69802624e-01 4.39309001e-01 2.53669530e-01\n",
" -6.71630800e-02 -2.37583295e-01 -5.01899600e-01 7.04075769e-02\n",
" -2.72109240e-01 5.25108771e-03]\n",
" [-3.17871541e-01 3.04395258e-01 -1.48960665e-01 1.34421825e-01\n",
" -1.49788372e-02 -2.33029246e-01 4.66970354e-01 6.02961361e-01\n",
" -1.70079693e-01 2.17916727e-01]\n",
" [ 3.32114726e-01 6.67006969e-02 -9.62992385e-02 -6.15109690e-02\n",
" 4.81428474e-01 -4.47902940e-02 1.71698347e-01 1.54229686e-01\n",
" -3.93822677e-02 5.50741851e-01]\n",
" [ 8.02836940e-02 1.29528269e-01 -1.31656677e-02 4.83411461e-01\n",
" -1.02488682e-01 -3.96515802e-02 8.13919865e-03 -1.78335369e-01\n",
" 5.28594971e-01 -5.06561756e-01]\n",
" [-3.25197518e-01 -4.00785357e-01 1.49951220e-01 3.13788801e-01\n",
" 1.76164478e-01 3.29165936e-01 -1.52645614e-02 -2.66170263e-01\n",
" 3.16257089e-01 1.78693980e-01]\n",
" [-4.10625517e-01 -4.73436117e-02 -4.05227207e-02 8.24426562e-02\n",
" 7.57036954e-02 -4.49555218e-01 -1.81903131e-02 7.56093115e-02\n",
" -1.58822700e-01 4.46603186e-02]], shape=(64, 10), dtype=float32)\n",
"\n",
"Example of one transformed row:\n",
"[ 0.22117373 -0.01721598 0.19548862 0.39414123 -0.0167326 -0.21437436\n",
" 0.1350686 -0.24686432 -0.29230666 0.11035113]\n"
]
}
],
"source": [
"demo(cat_embedded, categorical_tf_ds_batch)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.6"
}
},
"nbformat": 4,
"nbformat_minor": 4
}