Comparison tool refactor & ETL HUD RECAP (#272)

* Refactoring comparison tool and creating two new ETL notebooks

score/ipython/calenviroscreen_etl.ipynb

@@ -0,0 +1,141 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "20aa3891",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pathlib import Path\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import csv\n",
+    "import sys\n",
+    "import os\n",
+    "\n",
+    "module_path = os.path.abspath(os.path.join(\"..\"))\n",
+    "if module_path not in sys.path:\n",
+    "    sys.path.append(module_path)\n",
+    "\n",
+    "from etl.sources.census.etl_utils import get_state_fips_codes\n",
+    "from utils import unzip_file_from_url, remove_all_from_dir\n",
+    "\n",
+    "DATA_PATH = Path.cwd().parent / \"data\"\n",
+    "TMP_PATH = DATA_PATH / \"tmp\"\n",
+    "CALENVIROSCREEN_FTP_URL = \"https://justice40-data.s3.amazonaws.com/CalEnviroScreen/CalEnviroScreen_4.0_2021.zip\"\n",
+    "CSV_PATH = DATA_PATH / \"dataset\" / \"calenviroscreen4\"\n",
+    "\n",
+    "# Definining some variable names\n",
+    "CALENVIROSCREEN_SCORE_FIELD_NAME = \"calenviroscreen_score\"\n",
+    "CALENVIROSCREEN_PERCENTILE_FIELD_NAME = \"calenviroscreen_percentile\"\n",
+    "CALENVIROSCREEN_PRIORITY_COMMUNITY_FIELD_NAME = \"calenviroscreen_priority_community\"\n",
+    "GEOID_TRACT_FIELD_NAME = \"GEOID10_TRACT\"\n",
+    "\n",
+    "# Choosing constants.\n",
+    "# None of these numbers are final, but just for the purposes of comparison.\n",
+    "CALENVIROSCREEN_PRIORITY_COMMUNITY_THRESHOLD = 75\n",
+    "\n",
+    "print(DATA_PATH)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cc3fb9ec",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# download file from ejscreen ftp\n",
+    "unzip_file_from_url(CALENVIROSCREEN_FTP_URL, TMP_PATH, TMP_PATH)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "15f66756",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Data from https://calenviroscreen-oehha.hub.arcgis.com/#Data, specifically:\n",
+    "# https://oehha.ca.gov/media/downloads/calenviroscreen/document/calenviroscreen40resultsdatadictionaryd12021.zip\n",
+    "calenviroscreen_4_csv_name = \"CalEnviroScreen_4.0_2021.csv\"\n",
+    "calenviroscreen_data_path = TMP_PATH.joinpath(calenviroscreen_4_csv_name)\n",
+    "\n",
+    "# Load comparison index (CalEnviroScreen 4)\n",
+    "calenviroscreen_df = pd.read_csv(\n",
+    "    calenviroscreen_data_path, dtype={\"Census Tract\": \"string\"}\n",
+    ")\n",
+    "\n",
+    "calenviroscreen_df.rename(\n",
+    "    columns={\n",
+    "        \"Census Tract\": GEOID_TRACT_FIELD_NAME,\n",
+    "        \"DRAFT CES 4.0 Score\": CALENVIROSCREEN_SCORE_FIELD_NAME,\n",
+    "        \"DRAFT CES 4.0 Percentile\": CALENVIROSCREEN_PERCENTILE_FIELD_NAME,\n",
+    "    },\n",
+    "    inplace=True,\n",
+    ")\n",
+    "\n",
+    "# Add a leading \"0\" to the Census Tract to match our format in other data frames.\n",
+    "\n",
+    "calenviroscreen_df[GEOID_TRACT_FIELD_NAME] = (\n",
+    "    \"0\" + calenviroscreen_df[GEOID_TRACT_FIELD_NAME]\n",
+    ")\n",
+    "\n",
+    "# Calculate the top K% of prioritized communities\n",
+    "calenviroscreen_df[CALENVIROSCREEN_PRIORITY_COMMUNITY_FIELD_NAME] = (\n",
+    "    calenviroscreen_df[CALENVIROSCREEN_PERCENTILE_FIELD_NAME]\n",
+    "    >= CALENVIROSCREEN_PRIORITY_COMMUNITY_THRESHOLD\n",
+    ")\n",
+    "\n",
+    "calenviroscreen_df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9fa2077a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# write csv\n",
+    "CSV_PATH.mkdir(parents=True, exist_ok=True)\n",
+    "\n",
+    "# Matching other conventions in the ETL scripts, write only for the state (FIPS code 06).\n",
+    "calenviroscreen_df.to_csv(CSV_PATH / \"data06.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "81b977f8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# cleanup\n",
+    "remove_all_from_dir(TMP_PATH)"
+   ]
+  }
+ ],
+ "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": "3.7.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

score/ipython/hud_recap_etl.ipynb

@@ -0,0 +1,115 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "20aa3891",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pathlib import Path\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import csv\n",
+    "import sys\n",
+    "import os\n",
+    "\n",
+    "module_path = os.path.abspath(os.path.join(\"..\"))\n",
+    "if module_path not in sys.path:\n",
+    "    sys.path.append(module_path)\n",
+    "\n",
+    "from etl.sources.census.etl_utils import get_state_fips_codes\n",
+    "from utils import unzip_file_from_url, remove_all_from_dir\n",
+    "\n",
+    "DATA_PATH = Path.cwd().parent / \"data\"\n",
+    "TMP_PATH = DATA_PATH / \"tmp\"\n",
+    "HUD_RECAP_CSV_URL = \"https://opendata.arcgis.com/api/v3/datasets/56de4edea8264fe5a344da9811ef5d6e_0/downloads/data?format=csv&spatialRefId=4326\"\n",
+    "CSV_PATH = DATA_PATH / \"dataset\" / \"hud_recap\"\n",
+    "\n",
+    "# Definining some variable names\n",
+    "GEOID_TRACT_FIELD_NAME = \"GEOID10_TRACT\"\n",
+    "HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME = \"hud_recap_priority_community\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b9455da5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Data from https://hudgis-hud.opendata.arcgis.com/datasets/HUD::racially-or-ethnically-concentrated-areas-of-poverty-r-ecaps/about\n",
+    "df = pd.read_csv(HUD_RECAP_CSV_URL, dtype={\"GEOID\": \"string\"})\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ca63e66c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Rename some fields\n",
+    "df.rename(\n",
+    "    columns={\n",
+    "        \"GEOID\": GEOID_TRACT_FIELD_NAME,\n",
+    "        # Interestingly, there's no data dictionary for the RECAP data that I could find.\n",
+    "        # However, this site (http://www.schousing.com/library/Tax%20Credit/2020/QAP%20Instructions%20(2).pdf)\n",
+    "        # suggests:\n",
+    "        # \"If RCAP_Current for the tract in which the site is located is 1, the tract is an R/ECAP. If RCAP_Current is 0, it is not.\"\n",
+    "        \"RCAP_Current\": HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME,\n",
+    "    },\n",
+    "    inplace=True,\n",
+    ")\n",
+    "\n",
+    "# Convert to boolean\n",
+    "df[HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME] = df[\n",
+    "    HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME\n",
+    "].astype(\"bool\")\n",
+    "\n",
+    "df[HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME].value_counts()\n",
+    "\n",
+    "df.sort_values(by=GEOID_TRACT_FIELD_NAME, inplace=True)\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9fa2077a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# write csv\n",
+    "CSV_PATH.mkdir(parents=True, exist_ok=True)\n",
+    "\n",
+    "# Drop unnecessary columns.\n",
+    "df[[GEOID_TRACT_FIELD_NAME, HUD_RECAP_PRIORITY_COMMUNITY_FIELD_NAME]].to_csv(\n",
+    "    CSV_PATH / \"usa.csv\", index=False\n",
+    ")"
+   ]
+  }
+ ],
+ "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": "3.7.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

score/ipython/score_calc.ipynb

@@ -16,6 +16,7 @@
     "import collections\n",
     "import functools\n",
     "from pathlib import Path\n",
+    "import matplotlib.pyplot as plt\n",
     "import pandas as pd\n",
     "import csv\n",
     "import os\n",
@@ -363,7 +364,7 @@
    },
    "outputs": [],
    "source": [
-    "# calculate percentiles\n",
+    "# Calculate percentiles for each data set.\n",
     "for data_set in data_sets:\n",
     "    df[f\"{data_set.renamed_field}{PERCENTILE_FIELD_SUFFIX}\"] = df[\n",
     "        data_set.renamed_field\n",
@@ -379,7 +380,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# calculate min max\n",
+    "# Calculate min-max for each data set.\n",
     "# Math:\n",
     "# (\n",
     "#     Observed value\n",
@@ -410,6 +411,28 @@
     "df.head()"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f4eec326",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Graph distributions and correlations.\n",
+    "min_max_fields = [\n",
+    "    f\"{data_set.renamed_field}{MIN_MAX_FIELD_SUFFIX}\"\n",
+    "    for data_set in data_sets\n",
+    "    if data_set.renamed_field != GEOID_FIELD_NAME\n",
+    "]\n",
+    "df.hist(\n",
+    "    column=min_max_fields, layout=(len(min_max_fields), 1), figsize=(10, 30), bins=30\n",
+    ")\n",
+    "\n",
+    "plt.tight_layout()\n",
+    "\n",
+    "plt.show()"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -476,7 +499,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "fields_to_use_in_score = [\n",
+    "# Calculate scores D and E.\n",
+    "fields_to_use_in_score_d_and_e = [\n",
     "    UNEMPLOYED_FIELD_NAME,\n",
     "    LINGUISTIC_ISOLATION_FIELD_NAME,\n",
     "    HOUSING_BURDEN_FIELD_NAME,\n",
@@ -484,9 +508,11 @@
     "    HIGH_SCHOOL_FIELD_NAME,\n",
     "]\n",
     "\n",
-    "fields_min_max = [f\"{field}{MIN_MAX_FIELD_SUFFIX}\" for field in fields_to_use_in_score]\n",
+    "fields_min_max = [\n",
+    "    f\"{field}{MIN_MAX_FIELD_SUFFIX}\" for field in fields_to_use_in_score_d_and_e\n",
+    "]\n",
     "fields_percentile = [\n",
-    "    f\"{field}{PERCENTILE_FIELD_SUFFIX}\" for field in fields_to_use_in_score\n",
+    "    f\"{field}{PERCENTILE_FIELD_SUFFIX}\" for field in fields_to_use_in_score_d_and_e\n",
     "]\n",
     "\n",
     "# Calculate \"Score D\", which uses min-max normalization\n",
@@ -498,6 +524,32 @@
     "print(df[\"Score E\"].describe())"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a02e5bac",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Graph distributions\n",
+    "df.hist(\n",
+    "    column=fields_min_max, layout=(len(fields_min_max), 1), figsize=(10, 30), bins=30\n",
+    ")\n",
+    "plt.tight_layout()\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a0e608c8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Calculate correlations\n",
+    "df[fields_min_max].corr()"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,