Adding HOLC indicator (#1579)

Added HOLC indicator (Historic Redlining Score) from NCRC work; included 3.25 cutoff and low income as part of the housing burden category.

data/data-pipeline/data_pipeline/etl/constants.py

@@ -114,6 +114,11 @@ DATASET_LIST = [
         "module_dir": "maryland_ejscreen",
         "class_name": "MarylandEJScreenETL",
     },
+    {
+        "name": "historic_redlining",
+        "module_dir": "historic_redlining",
+        "class_name": "HistoricRedliningETL",
+    },
     # This has to come after us.json exists
     {
         "name": "census_acs",

data/data-pipeline/data_pipeline/etl/score/constants.py

@@ -205,7 +205,8 @@ TILES_SCORE_COLUMNS = {
     field_names.M_HEALTH: "M_HLTH",
     # temporarily update this so that it's the Narwhal score that gets visualized on the map
     field_names.SCORE_N_COMMUNITIES: "SM_C",
-    field_names.SCORE_M + field_names.PERCENTILE_FIELD_SUFFIX: "SM_PFS",
+    field_names.SCORE_N_COMMUNITIES
+    + field_names.PERCENTILE_FIELD_SUFFIX: "SM_PFS",
     field_names.EXPECTED_POPULATION_LOSS_RATE_LOW_INCOME_LOW_HIGHER_ED_FIELD: "EPLRLI",
     field_names.EXPECTED_AGRICULTURE_LOSS_RATE_LOW_INCOME_LOW_HIGHER_ED_FIELD: "EALRLI",
     field_names.EXPECTED_BUILDING_LOSS_RATE_LOW_INCOME_LOW_HIGHER_ED_FIELD: "EBLRLI",

data/data-pipeline/data_pipeline/etl/score/etl_score.py

@@ -1,5 +1,6 @@
 import functools
 from collections import namedtuple
+from attr import field
 
 import numpy as np
 import pandas as pd
@@ -36,6 +37,7 @@ class ScoreETL(ExtractTransformLoad):
         self.census_decennial_df: pd.DataFrame
         self.census_2010_df: pd.DataFrame
         self.child_opportunity_index_df: pd.DataFrame
+        self.hrs_df: pd.DataFrame
 
     def extract(self) -> None:
         logger.info("Loading data sets from disk.")
@@ -172,6 +174,17 @@ class ScoreETL(ExtractTransformLoad):
             low_memory=False,
         )
 
+        # Load HRS data
+        hrs_csv = (
+            constants.DATA_PATH / "dataset" / "historic_redlining" / "usa.csv"
+        )
+
+        self.hrs_df = pd.read_csv(
+            hrs_csv,
+            dtype={self.GEOID_TRACT_FIELD_NAME: "string"},
+            low_memory=False,
+        )
+
     def _join_tract_dfs(self, census_tract_dfs: list) -> pd.DataFrame:
         logger.info("Joining Census Tract dataframes")
 
@@ -376,6 +389,7 @@ class ScoreETL(ExtractTransformLoad):
             self.census_decennial_df,
             self.census_2010_df,
             self.child_opportunity_index_df,
+            self.hrs_df,
         ]
 
         # Sanity check each data frame before merging.
@@ -405,7 +419,6 @@ class ScoreETL(ExtractTransformLoad):
             df[field_names.MEDIAN_INCOME_FIELD] / df[field_names.AMI_FIELD]
         )
 
-        # QQ: why don't we just filter to the numeric columns by type?
         numeric_columns = [
             field_names.HOUSING_BURDEN_FIELD,
             field_names.TOTAL_POP_FIELD,
@@ -465,6 +478,7 @@ class ScoreETL(ExtractTransformLoad):
         non_numeric_columns = [
             self.GEOID_TRACT_FIELD_NAME,
             field_names.PERSISTENT_POVERTY_FIELD,
+            field_names.HISTORIC_REDLINING_SCORE_EXCEEDED,
         ]
 
         # For some columns, high values are "good", so we want to reverse the percentile

data/data-pipeline/data_pipeline/etl/score/etl_score_geo.py

@@ -46,10 +46,11 @@ class GeoScoreETL(ExtractTransformLoad):
             self.DATA_PATH / "census" / "geojson" / "us.json"
         )
 
-        # Import the shortened name for Score M percentile ("SM_PFS") that's used on the
+        # Import the shortened name for Score N percentile ("SM_PFS") that's used on the
         # tiles.
+        ## TEMPORARY update
         self.TARGET_SCORE_SHORT_FIELD = constants.TILES_SCORE_COLUMNS[
-            field_names.SCORE_M + field_names.PERCENTILE_FIELD_SUFFIX
+            field_names.SCORE_N + field_names.PERCENTILE_FIELD_SUFFIX
         ]
         self.TARGET_SCORE_RENAME_TO = "M_SCORE"
 
@@ -284,21 +285,28 @@ class GeoScoreETL(ExtractTransformLoad):
 
         def create_esri_codebook(codebook):
             """temporary: helper to make a codebook for esri shapefile only"""
+<<<<<<< HEAD
 
             shapefile_column_field = "shapefile_column"
             internal_column_name_field = "column_name"
             column_description_field = "column_description"
 
+=======
+>>>>>>> 8c255f0e (Adding HOLC indicator (#1579))
             logger.info("Creating a codebook that uses the csv names")
             codebook = (
                 pd.Series(codebook)
                 .reset_index()
                 .rename(
                     # kept as strings because no downstream impacts
+<<<<<<< HEAD
                     columns={
                         0: internal_column_name_field,
                         "index": shapefile_column_field,
                     }
+=======
+                    columns={0: "column_name", "index": "shapefile_column"}
+>>>>>>> 8c255f0e (Adding HOLC indicator (#1579))
                 )
             )
 
@@ -374,7 +382,7 @@ class GeoScoreETL(ExtractTransformLoad):
                 for task in [
                     write_high_to_file,
                     write_low_to_file,
-                    write_esri_shapefile,
+                    # write_esri_shapefile,
                 ]
             }
 

data/data-pipeline/data_pipeline/etl/sources/historic_redlining/etl.py

@@ -0,0 +1,72 @@
+import pandas as pd
+
+from data_pipeline.etl.base import ExtractTransformLoad
+from data_pipeline.utils import get_module_logger
+from data_pipeline.config import settings
+
+logger = get_module_logger(__name__)
+
+
+class HistoricRedliningETL(ExtractTransformLoad):
+    def __init__(self):
+        self.CSV_PATH = self.DATA_PATH / "dataset" / "historic_redlining"
+        self.HISTORIC_REDLINING_URL = (
+            settings.AWS_JUSTICE40_DATASOURCES_URL + "/HRS_2010.zip"
+        )
+        self.HISTORIC_REDLINING_FILE_PATH = (
+            self.get_tmp_path() / "HRS_2010.xlsx"
+        )
+
+        self.REDLINING_SCALAR = "Tract-level redlining score"
+
+        self.COLUMNS_TO_KEEP = [
+            self.GEOID_TRACT_FIELD_NAME,
+            self.REDLINING_SCALAR,
+        ]
+        self.df: pd.DataFrame
+
+    def extract(self) -> None:
+        logger.info("Downloading Historic Redlining Data")
+        super().extract(
+            self.HISTORIC_REDLINING_URL,
+            self.get_tmp_path(),
+        )
+
+    def transform(self) -> None:
+        logger.info("Transforming Historic Redlining Data")
+        # this is obviously temporary
+        historic_redlining_data = pd.read_excel(
+            self.HISTORIC_REDLINING_FILE_PATH
+        )
+        historic_redlining_data[self.GEOID_TRACT_FIELD_NAME] = (
+            historic_redlining_data["GEOID10"].astype(str).str.zfill(11)
+        )
+        historic_redlining_data = historic_redlining_data.rename(
+            columns={"HRS2010": self.REDLINING_SCALAR}
+        )
+
+        logger.info(f"{historic_redlining_data.columns}")
+
+        # Calculate lots of different score thresholds for convenience
+        for threshold in [3.25, 3.5, 3.75]:
+            historic_redlining_data[
+                f"{self.REDLINING_SCALAR} meets or exceeds {round(threshold, 2)}"
+            ] = (historic_redlining_data[self.REDLINING_SCALAR] >= threshold)
+            ## NOTE We add to columns to keep here
+            self.COLUMNS_TO_KEEP.append(
+                f"{self.REDLINING_SCALAR} meets or exceeds {round(threshold, 2)}"
+            )
+
+        self.df = historic_redlining_data
+
+    def load(self) -> None:
+        logger.info("Saving Historic Redlining CSV")
+        # write selected states csv
+        self.CSV_PATH.mkdir(parents=True, exist_ok=True)
+        self.df[self.COLUMNS_TO_KEEP].to_csv(
+            self.CSV_PATH / "usa.csv", index=False
+        )
+
+    def validate(self) -> None:
+        logger.info("Validating Historic Redlining Data")
+        pass

data/data-pipeline/data_pipeline/etl/sources/mapping_inequality/etl.py

@@ -47,16 +47,21 @@ class MappingInequalityETL(ExtractTransformLoad):
         self.HOLC_GRADE_AND_ID_FIELD: str = "holc_id"
         self.CITY_INPUT_FIELD: str = "city"
 
-        self.HOLC_GRADE_D_FIELD: str = "HOLC Grade D"
+        self.HOLC_GRADE_D_FIELD: str = "HOLC Grade D (hazardous)"
+        self.HOLC_GRADE_C_FIELD: str = "HOLC Grade C (declining)"
         self.HOLC_GRADE_MANUAL_FIELD: str = "HOLC Grade (manually mapped)"
         self.HOLC_GRADE_DERIVED_FIELD: str = "HOLC Grade (derived)"
 
         self.COLUMNS_TO_KEEP = [
             self.GEOID_TRACT_FIELD_NAME,
+            field_names.HOLC_GRADE_C_TRACT_PERCENT_FIELD,
+            field_names.HOLC_GRADE_C_OR_D_TRACT_PERCENT_FIELD,
+            field_names.HOLC_GRADE_C_OR_D_TRACT_50_PERCENT_FIELD,
             field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD,
             field_names.HOLC_GRADE_D_TRACT_20_PERCENT_FIELD,
             field_names.HOLC_GRADE_D_TRACT_50_PERCENT_FIELD,
             field_names.HOLC_GRADE_D_TRACT_75_PERCENT_FIELD,
+            field_names.REDLINED_SHARE,
         ]
 
         self.df: pd.DataFrame
@@ -113,34 +118,58 @@ class MappingInequalityETL(ExtractTransformLoad):
             how="left",
         )
 
-        # Create a single field that combines the 'derived' grade D field with the
+        # Create a single field that combines the 'derived' grade C and D fields with the
-        # manually mapped grade D field into a single grade D field.
+        # manually mapped grade C and D field into a single grade C and D field.
-        merged_df[self.HOLC_GRADE_D_FIELD] = np.where(
+        ## Note: there are no manually derived C tracts at the moment
-            (merged_df[self.HOLC_GRADE_DERIVED_FIELD] == "D")
-            | (merged_df[self.HOLC_GRADE_MANUAL_FIELD] == "D"),
-            True,
-            None,
-        )
 
-        # Start grouping by, to sum all of the grade D parts of each tract.
+        for grade, field_name in [
-        grouped_df = (
+            ("C", self.HOLC_GRADE_C_FIELD),
-            merged_df.groupby(
+            ("D", self.HOLC_GRADE_D_FIELD),
-                by=[
+        ]:
-                    self.GEOID_TRACT_FIELD_NAME,
+            merged_df[field_name] = np.where(
-                    self.HOLC_GRADE_D_FIELD,
+                (merged_df[self.HOLC_GRADE_DERIVED_FIELD] == grade)
-                ],
+                | (merged_df[self.HOLC_GRADE_MANUAL_FIELD] == grade),
-                # Keep the nulls, so we know the non-D proportion.
+                True,
-                dropna=False,
+                None,
-            )[self.TRACT_PROPORTION_FIELD]
+            )
+
+        redlined_dataframes_list = [
+            merged_df[merged_df[field].fillna(False)]
+            .groupby(self.GEOID_TRACT_FIELD_NAME)[self.TRACT_PROPORTION_FIELD]
             .sum()
+            .rename(new_name)
+            for field, new_name in [
+                (
+                    self.HOLC_GRADE_D_FIELD,
+                    field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD,
+                ),
+                (
+                    self.HOLC_GRADE_C_FIELD,
+                    field_names.HOLC_GRADE_C_TRACT_PERCENT_FIELD,
+                ),
+            ]
+        ]
+
+        # Group by tract ID to get tract proportions of just C or just D
+        # This produces a single row per tract
+        grouped_df = (
+            pd.concat(
+                redlined_dataframes_list,
+                axis=1,
+            )
+            .fillna(0)
             .reset_index()
         )
 
-        # Create a field that is only the percent that is grade D.
+        grouped_df[
-        grouped_df[field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD] = np.where(
+            field_names.HOLC_GRADE_C_OR_D_TRACT_PERCENT_FIELD
-            grouped_df[self.HOLC_GRADE_D_FIELD],
+        ] = grouped_df[
-            grouped_df[self.TRACT_PROPORTION_FIELD],
+            [
-            0,
+                field_names.HOLC_GRADE_C_TRACT_PERCENT_FIELD,
+                field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD,
+            ]
+        ].sum(
+            axis=1
         )
 
         # Calculate some specific threshold cutoffs, for convenience.
@@ -154,15 +183,14 @@ class MappingInequalityETL(ExtractTransformLoad):
             grouped_df[field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD] > 0.75
         )
 
-        # Drop the non-True values of `self.HOLC_GRADE_D_FIELD` -- we only
+        grouped_df[field_names.HOLC_GRADE_C_OR_D_TRACT_50_PERCENT_FIELD] = (
-        # want one row per tract for future joins.
+            grouped_df[field_names.HOLC_GRADE_C_OR_D_TRACT_PERCENT_FIELD] > 0.5
-        # Note this means not all tracts will be in this data.
+        )
-        # Note: this singleton comparison warning may be a pylint bug:
+
-        # https://stackoverflow.com/questions/51657715/pylint-pandas-comparison-to-true-should-be-just-expr-or-expr-is-true-sin#comment90876517_51657715
+        # Create the indicator we will use
-        # pylint: disable=singleton-comparison
+        grouped_df[field_names.REDLINED_SHARE] = (
-        grouped_df = grouped_df[
+            grouped_df[field_names.HOLC_GRADE_C_OR_D_TRACT_PERCENT_FIELD] > 0.5
-            grouped_df[self.HOLC_GRADE_D_FIELD] == True  # noqa: E712
+        ) & (grouped_df[field_names.HOLC_GRADE_D_TRACT_PERCENT_FIELD] > 0)
-        ]
 
         # Sort for convenience.
         grouped_df.sort_values(by=self.GEOID_TRACT_FIELD_NAME, inplace=True)

data/data-pipeline/data_pipeline/score/field_names.py

@@ -57,7 +57,7 @@ M_WORKFORCE = "Workforce Factor (Definition M)"
 M_NON_WORKFORCE = "Any Non-Workforce Factor (Definition M)"
 
 # Definition Narwhal fields
-SCORE_N = "Definition N"
+SCORE_N = "Definition N (communities)"
 SCORE_N_COMMUNITIES = "Definition N (communities)"
 N_CLIMATE = "Climate Factor (Definition N)"
 N_ENERGY = "Energy Factor (Definition N)"
@@ -303,7 +303,17 @@ EJSCREEN_AREAS_OF_CONCERN_STATE_95TH_PERCENTILE_COMMUNITIES_FIELD = (
     "EJSCREEN Areas of Concern, State, 95th percentile (communities)"
 )
 # Mapping inequality data.
+REDLINED_SHARE: str = (
+    "Redlined share: tract had redlining and was more than 50% Grade C or D"
+)
 HOLC_GRADE_D_TRACT_PERCENT_FIELD: str = "Percent of tract that is HOLC Grade D"
+HOLC_GRADE_C_TRACT_PERCENT_FIELD: str = "Percent of tract that is HOLC Grade C"
+HOLC_GRADE_C_OR_D_TRACT_PERCENT_FIELD: str = (
+    "Percent of tract that is HOLC Grade C or HOLC Grade D"
+)
+HOLC_GRADE_C_OR_D_TRACT_50_PERCENT_FIELD: str = (
+    "Tract is more than 50% Grade C or D"
+)
 HOLC_GRADE_D_TRACT_20_PERCENT_FIELD: str = "Tract is >20% HOLC Grade D"
 HOLC_GRADE_D_TRACT_50_PERCENT_FIELD: str = "Tract is >50% HOLC Grade D"
 HOLC_GRADE_D_TRACT_75_PERCENT_FIELD: str = "Tract is >75% HOLC Grade D"
@@ -316,7 +326,7 @@ MICHIGAN_EJSCREEN_PRIORITY_COMMUNITY_FIELD: str = (
 )
 
 # CDC SVI INDEX percentile fields
-CDC_SVI_INDEX_SE_THEME_FIELD: str = "SVI - Socioeconomic Index"
+CDC_SVI_INDEX_SE_THEME_FIELD: str = "SVI - Social Vulnerability Index"
 CDC_SVI_INDEX_HOUSEHOLD_THEME_COMPOSITION_FIELD: str = (
     "SVI - Household Composition Index"
 )
@@ -691,5 +701,14 @@ MAPPING_FOR_EJ_PRIORITY_COMMUNITY_FIELD = (
     "Mapping for Environmental Justice Priority Community"
 )
 
+# Historic Redlining Score
+HISTORIC_REDLINING_SCORE_EXCEEDED = (
+    "Tract-level redlining score meets or exceeds 3.25"
+)
+
+HISTORIC_REDLINING_SCORE_EXCEEDED_LOW_INCOME_FIELD = (
+    "Tract-level redlining score meets or exceeds 3.25 and is low income"
+)
+
 # End of names for individual factors being exceeded
 ####

data/data-pipeline/data_pipeline/score/score_narwhal.py

@@ -1,4 +1,5 @@
 from typing import Tuple
+from attr import field
 import numpy as np
 import pandas as pd
 
@@ -308,11 +309,22 @@ class ScoreNarwhal(Score):
         # poverty level and has a low percent of higher ed students.
         # Source: Census's American Community Survey
 
+        ## Additionally, we look to see if HISTORIC_REDLINING_SCORE_EXCEEDED is True and the tract is also low income
+
         housing_eligibility_columns = [
             field_names.LEAD_PAINT_MEDIAN_HOUSE_VALUE_LOW_INCOME_FIELD,
             field_names.HOUSING_BURDEN_LOW_INCOME_FIELD,
+            field_names.HISTORIC_REDLINING_SCORE_EXCEEDED_LOW_INCOME_FIELD,
         ]
 
+        # design question -- should read in scalar with threshold here instead?
+        self.df[
+            field_names.HISTORIC_REDLINING_SCORE_EXCEEDED_LOW_INCOME_FIELD
+        ] = (
+            self.df[field_names.HISTORIC_REDLINING_SCORE_EXCEEDED]
+            & self.df[field_names.FPL_200_SERIES_IMPUTED_AND_ADJUSTED]
+        )
+
         self.df[field_names.LEAD_PAINT_PROXY_PCTILE_THRESHOLD] = (
             self.df[
                 field_names.LEAD_PAINT_FIELD
@@ -804,5 +816,8 @@ class ScoreNarwhal(Score):
         ]
         self.df[field_names.CATEGORY_COUNT] = self.df[factors].sum(axis=1)
         self.df[field_names.SCORE_N_COMMUNITIES] = self.df[factors].any(axis=1)
+        self.df[
+            field_names.SCORE_N_COMMUNITIES + field_names.PERCENTILE_FIELD_SUFFIX
+        ] = self.df[field_names.SCORE_N_COMMUNITIES].astype(int)
 
         return self.df