publicdata/j40-cejst-2
1
0
Fork 0
mirror of https://github.com/DOI-DO/j40-cejst-2.git synced 2025-07-30 12:21:17 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

Data directory should adopt standard Poetry-suggested python package structure (#457)

Browse source 
* Fixes #456 - Our data directory should adopt standard python package structure
* a few missed references
* updating readme
* updating requirements
* Running Black
* Fixes for flake8
* updating pylint
This commit is contained in:
Nat Hillard 2021-08-05 15:35:54 -04:00 • committed by GitHub
commit c1568e87c0
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
61 changed files with 1273 additions and 1256 deletions
Download patch file Download diff file Expand all files Collapse all files

0
data/data-pipeline/data_pipeline/etl/score/__init__.py Normal file
View file

401
data/data-pipeline/data_pipeline/etl/score/etl_score.py Normal file
View file

@ -0,0 +1,401 @@
import collections
import functools
import pandas as pd
from data_pipeline.etl.base import ExtractTransformLoad
from data_pipeline.utils import get_module_logger
logger = get_module_logger(__name__)
class ScoreETL(ExtractTransformLoad):
def __init__(self):
# Define some global parameters
self.BUCKET_SOCIOECONOMIC = "Socioeconomic Factors"
self.BUCKET_SENSITIVE = "Sensitive populations"
self.BUCKET_ENVIRONMENTAL = "Environmental effects"
self.BUCKET_EXPOSURES = "Exposures"
self.BUCKETS = [
self.BUCKET_SOCIOECONOMIC,
self.BUCKET_SENSITIVE,
self.BUCKET_ENVIRONMENTAL,
self.BUCKET_EXPOSURES,
]
# A few specific field names
# TODO: clean this up, I name some fields but not others.
self.UNEMPLOYED_FIELD_NAME = "Unemployed civilians (percent)"
self.LINGUISTIC_ISOLATION_FIELD_NAME = "Linguistic isolation (percent)"
self.HOUSING_BURDEN_FIELD_NAME = "Housing burden (percent)"
self.POVERTY_FIELD_NAME = "Poverty (Less than 200% of federal poverty line)"
self.HIGH_SCHOOL_FIELD_NAME = (
"Percent individuals age 25 or over with less than high school degree"
)
# There's another aggregation level (a second level of "buckets").
self.AGGREGATION_POLLUTION = "Pollution Burden"
self.AGGREGATION_POPULATION = "Population Characteristics"
self.PERCENTILE_FIELD_SUFFIX = " (percentile)"
self.MIN_MAX_FIELD_SUFFIX = " (min-max normalized)"
self.SCORE_CSV_PATH = self.DATA_PATH / "score" / "csv" / "full"
# dataframes
self.df: pd.DataFrame
self.ejscreen_df: pd.DataFrame
self.census_df: pd.DataFrame
self.housing_and_transportation_df: pd.DataFrame
self.hud_housing_df: pd.DataFrame
def extract(self) -> None:
# EJSCreen csv Load
ejscreen_csv = self.DATA_PATH / "dataset" / "ejscreen_2019" / "usa.csv"
self.ejscreen_df = pd.read_csv(
ejscreen_csv, dtype={"ID": "string"}, low_memory=False
)
self.ejscreen_df.rename(columns={"ID": self.GEOID_FIELD_NAME}, inplace=True)
# Load census data
census_csv = self.DATA_PATH / "dataset" / "census_acs_2019" / "usa.csv"
self.census_df = pd.read_csv(
census_csv, dtype={self.GEOID_FIELD_NAME: "string"}, low_memory=False,
)
# Load housing and transportation data
housing_and_transportation_index_csv = (
self.DATA_PATH / "dataset" / "housing_and_transportation_index" / "usa.csv"
)
self.housing_and_transportation_df = pd.read_csv(
housing_and_transportation_index_csv,
dtype={self.GEOID_FIELD_NAME: "string"},
low_memory=False,
)
# Load HUD housing data
hud_housing_csv = self.DATA_PATH / "dataset" / "hud_housing" / "usa.csv"
self.hud_housing_df = pd.read_csv(
hud_housing_csv,
dtype={self.GEOID_TRACT_FIELD_NAME: "string"},
low_memory=False,
)
def transform(self) -> None:
logger.info("Transforming Score Data")
# Join all the data sources that use census block groups
census_block_group_dfs = [
self.ejscreen_df,
self.census_df,
self.housing_and_transportation_df,
]
census_block_group_df = functools.reduce(
lambda left, right: pd.merge(
left=left, right=right, on=self.GEOID_FIELD_NAME, how="outer"
),
census_block_group_dfs,
)
# Sanity check the join.
if len(census_block_group_df[self.GEOID_FIELD_NAME].str.len().unique()) != 1:
raise ValueError(
f"One of the input CSVs uses {self.GEOID_FIELD_NAME} with a different length."
)
# Join all the data sources that use census tracts
# TODO: when there's more than one data source using census tract, reduce/merge them here.
census_tract_df = self.hud_housing_df
# Calculate the tract for the CBG data.
census_block_group_df[self.GEOID_TRACT_FIELD_NAME] = census_block_group_df[
self.GEOID_FIELD_NAME
].str[0:11]
self.df = census_block_group_df.merge(
census_tract_df, on=self.GEOID_TRACT_FIELD_NAME
)
if len(census_block_group_df) > 220333:
raise ValueError("Too many rows in the join.")
# Define a named tuple that will be used for each data set input.
DataSet = collections.namedtuple(
typename="DataSet", field_names=["input_field", "renamed_field", "bucket"],
)
data_sets = [
# The following data sets have `bucket=None`, because it's not used in the bucket based score ("Score C").
DataSet(
input_field=self.GEOID_FIELD_NAME,
# Use the name `GEOID10` to enable geoplatform.gov's workflow.
renamed_field=self.GEOID_FIELD_NAME,
bucket=None,
),
DataSet(
input_field=self.HOUSING_BURDEN_FIELD_NAME,
renamed_field=self.HOUSING_BURDEN_FIELD_NAME,
bucket=None,
),
DataSet(
input_field="ACSTOTPOP", renamed_field="Total population", bucket=None,
),
# The following data sets have buckets, because they're used in the score
DataSet(
input_field="CANCER",
renamed_field="Air toxics cancer risk",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="RESP",
renamed_field="Respiratory hazard index",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="DSLPM",
renamed_field="Diesel particulate matter",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="PM25",
renamed_field="Particulate matter (PM2.5)",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="OZONE",
renamed_field="Ozone",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="PTRAF",
renamed_field="Traffic proximity and volume",
bucket=self.BUCKET_EXPOSURES,
),
DataSet(
input_field="PRMP",
renamed_field="Proximity to RMP sites",
bucket=self.BUCKET_ENVIRONMENTAL,
),
DataSet(
input_field="PTSDF",
renamed_field="Proximity to TSDF sites",
bucket=self.BUCKET_ENVIRONMENTAL,
),
DataSet(
input_field="PNPL",
renamed_field="Proximity to NPL sites",
bucket=self.BUCKET_ENVIRONMENTAL,
),
DataSet(
input_field="PWDIS",
renamed_field="Wastewater discharge",
bucket=self.BUCKET_ENVIRONMENTAL,
),
DataSet(
input_field="PRE1960PCT",
renamed_field="Percent pre-1960s housing (lead paint indicator)",
bucket=self.BUCKET_ENVIRONMENTAL,
),
DataSet(
input_field="UNDER5PCT",
renamed_field="Individuals under 5 years old",
bucket=self.BUCKET_SENSITIVE,
),
DataSet(
input_field="OVER64PCT",
renamed_field="Individuals over 64 years old",
bucket=self.BUCKET_SENSITIVE,
),
DataSet(
input_field=self.LINGUISTIC_ISOLATION_FIELD_NAME,
renamed_field=self.LINGUISTIC_ISOLATION_FIELD_NAME,
bucket=self.BUCKET_SENSITIVE,
),
DataSet(
input_field="LINGISOPCT",
renamed_field="Percent of households in linguistic isolation",
bucket=self.BUCKET_SOCIOECONOMIC,
),
DataSet(
input_field="LOWINCPCT",
renamed_field=self.POVERTY_FIELD_NAME,
bucket=self.BUCKET_SOCIOECONOMIC,
),
DataSet(
input_field="LESSHSPCT",
renamed_field=self.HIGH_SCHOOL_FIELD_NAME,
bucket=self.BUCKET_SOCIOECONOMIC,
),
DataSet(
input_field=self.UNEMPLOYED_FIELD_NAME,
renamed_field=self.UNEMPLOYED_FIELD_NAME,
bucket=self.BUCKET_SOCIOECONOMIC,
),
DataSet(
input_field="ht_ami",
renamed_field="Housing + Transportation Costs % Income for the Regional Typical Household",
bucket=self.BUCKET_SOCIOECONOMIC,
),
]
# Rename columns:
renaming_dict = {
data_set.input_field: data_set.renamed_field for data_set in data_sets
}
self.df.rename(
columns=renaming_dict, inplace=True, errors="raise",
)
columns_to_keep = [data_set.renamed_field for data_set in data_sets]
self.df = self.df[columns_to_keep]
# Convert all columns to numeric.
for data_set in data_sets:
# Skip GEOID_FIELD_NAME, because it's a string.
if data_set.renamed_field == self.GEOID_FIELD_NAME:
continue
self.df[f"{data_set.renamed_field}"] = pd.to_numeric(
self.df[data_set.renamed_field]
)
# calculate percentiles
for data_set in data_sets:
self.df[
f"{data_set.renamed_field}{self.PERCENTILE_FIELD_SUFFIX}"
] = self.df[data_set.renamed_field].rank(pct=True)
# Math:
# (
# Observed value
# - minimum of all values
# )
# divided by
# (
# Maximum of all values
# - minimum of all values
# )
for data_set in data_sets:
# Skip GEOID_FIELD_NAME, because it's a string.
if data_set.renamed_field == self.GEOID_FIELD_NAME:
continue
min_value = self.df[data_set.renamed_field].min(skipna=True)
max_value = self.df[data_set.renamed_field].max(skipna=True)
logger.info(
f"For data set {data_set.renamed_field}, the min value is {min_value} and the max value is {max_value}."
)
self.df[f"{data_set.renamed_field}{self.MIN_MAX_FIELD_SUFFIX}"] = (
self.df[data_set.renamed_field] - min_value
) / (max_value - min_value)
# Graph distributions and correlations.
min_max_fields = [ # noqa: F841
f"{data_set.renamed_field}{self.MIN_MAX_FIELD_SUFFIX}"
for data_set in data_sets
if data_set.renamed_field != self.GEOID_FIELD_NAME
]
# Calculate score "A" and score "B"
self.df["Score A"] = self.df[
[
"Poverty (Less than 200% of federal poverty line) (percentile)",
"Percent individuals age 25 or over with less than high school degree (percentile)",
]
].mean(axis=1)
self.df["Score B"] = (
self.df["Poverty (Less than 200% of federal poverty line) (percentile)"]
* self.df[
"Percent individuals age 25 or over with less than high school degree (percentile)"
]
)
# Calculate "CalEnviroScreen for the US" score
# Average all the percentile values in each bucket into a single score for each of the four buckets.
for bucket in self.BUCKETS:
fields_in_bucket = [
f"{data_set.renamed_field}{self.PERCENTILE_FIELD_SUFFIX}"
for data_set in data_sets
if data_set.bucket == bucket
]
self.df[f"{bucket}"] = self.df[fields_in_bucket].mean(axis=1)
# Combine the score from the two Exposures and Environmental Effects buckets
# into a single score called "Pollution Burden".
# The math for this score is:
# (1.0 * Exposures Score + 0.5 * Environment Effects score) / 1.5.
self.df[self.AGGREGATION_POLLUTION] = (
1.0 * self.df[f"{self.BUCKET_EXPOSURES}"]
+ 0.5 * self.df[f"{self.BUCKET_ENVIRONMENTAL}"]
) / 1.5
# Average the score from the two Sensitive populations and
# Socioeconomic factors buckets into a single score called
# "Population Characteristics".
self.df[self.AGGREGATION_POPULATION] = self.df[
[f"{self.BUCKET_SENSITIVE}", f"{self.BUCKET_SOCIOECONOMIC}"]
].mean(axis=1)
# Multiply the "Pollution Burden" score and the "Population Characteristics"
# together to produce the cumulative impact score.
self.df["Score C"] = (
self.df[self.AGGREGATION_POLLUTION] * self.df[self.AGGREGATION_POPULATION]
)
if len(census_block_group_df) > 220333:
raise ValueError("Too many rows in the join.")
fields_to_use_in_score = [
self.UNEMPLOYED_FIELD_NAME,
self.LINGUISTIC_ISOLATION_FIELD_NAME,
self.HOUSING_BURDEN_FIELD_NAME,
self.POVERTY_FIELD_NAME,
self.HIGH_SCHOOL_FIELD_NAME,
]
fields_min_max = [
f"{field}{self.MIN_MAX_FIELD_SUFFIX}" for field in fields_to_use_in_score
]
fields_percentile = [
f"{field}{self.PERCENTILE_FIELD_SUFFIX}" for field in fields_to_use_in_score
]
# Calculate "Score D", which uses min-max normalization
# and calculate "Score E", which uses percentile normalization for the same fields
self.df["Score D"] = self.df[fields_min_max].mean(axis=1)
self.df["Score E"] = self.df[fields_percentile].mean(axis=1)
# Calculate correlations
self.df[fields_min_max].corr()
# Create percentiles for the scores
for score_field in [
"Score A",
"Score B",
"Score C",
"Score D",
"Score E",
"Poverty (Less than 200% of federal poverty line)",
]:
self.df[f"{score_field}{self.PERCENTILE_FIELD_SUFFIX}"] = self.df[
score_field
].rank(pct=True)
for threshold in [0.25, 0.3, 0.35, 0.4]:
fraction_converted_to_percent = int(100 * threshold)
self.df[
f"{score_field} (top {fraction_converted_to_percent}th percentile)"
] = (
self.df[f"{score_field}{self.PERCENTILE_FIELD_SUFFIX}"]
>= 1 - threshold
)
def load(self) -> None:
logger.info("Saving Score CSV")
# write nationwide csv
self.SCORE_CSV_PATH.mkdir(parents=True, exist_ok=True)
self.df.to_csv(self.SCORE_CSV_PATH / "usa.csv", index=False)

156
data/data-pipeline/data_pipeline/etl/score/etl_score_geo.py Normal file
View file

@ -0,0 +1,156 @@
import math
import pandas as pd
import geopandas as gpd
from data_pipeline.etl.base import ExtractTransformLoad
from data_pipeline.utils import get_module_logger
logger = get_module_logger(__name__)
class GeoScoreETL(ExtractTransformLoad):
"""
A class used to generate per state and national GeoJson files with the score baked in
"""
def __init__(self):
self.SCORE_GEOJSON_PATH = self.DATA_PATH / "score" / "geojson"
self.SCORE_LOW_GEOJSON = self.SCORE_GEOJSON_PATH / "usa-low.json"
self.SCORE_HIGH_GEOJSON = self.SCORE_GEOJSON_PATH / "usa-high.json"
self.SCORE_CSV_PATH = self.DATA_PATH / "score" / "csv"
self.TILE_SCORE_CSV = self.SCORE_CSV_PATH / "tiles" / "usa.csv"
self.CENSUS_USA_GEOJSON = self.DATA_PATH / "census" / "geojson" / "us.json"
self.TARGET_SCORE_NAME = "Score E (percentile)"
self.TARGET_SCORE_RENAME_TO = "E_SCORE"
self.NUMBER_OF_BUCKETS = 10
self.geojson_usa_df: gpd.GeoDataFrame
self.score_usa_df: pd.DataFrame
self.geojson_score_usa_high: gpd.GeoDataFrame
self.geojson_score_usa_low: gpd.GeoDataFrame
def extract(self) -> None:
logger.info("Reading US GeoJSON (~6 minutes)")
self.geojson_usa_df = gpd.read_file(
self.CENSUS_USA_GEOJSON,
dtype={"GEOID10": "string"},
usecols=["GEOID10", "geometry"],
low_memory=False,
)
self.geojson_usa_df.head()
logger.info("Reading score CSV")
self.score_usa_df = pd.read_csv(
self.TILE_SCORE_CSV, dtype={"GEOID10": "string"}, low_memory=False,
)
def transform(self) -> None:
logger.info("Pruning Census GeoJSON")
fields = ["GEOID10", "geometry"]
self.geojson_usa_df = self.geojson_usa_df[fields]
logger.info("Merging and compressing score CSV with USA GeoJSON")
self.geojson_score_usa_high = self.score_usa_df.merge(
self.geojson_usa_df, on="GEOID10", how="left"
)
self.geojson_score_usa_high = gpd.GeoDataFrame(
self.geojson_score_usa_high, crs="EPSG:4326"
)
usa_simplified = self.geojson_score_usa_high[
["GEOID10", self.TARGET_SCORE_NAME, "geometry"]
].reset_index(drop=True)
usa_simplified.rename(
columns={self.TARGET_SCORE_NAME: self.TARGET_SCORE_RENAME_TO}, inplace=True,
)
logger.info("Aggregating into tracts (~5 minutes)")
usa_tracts = self._aggregate_to_tracts(usa_simplified)
usa_tracts = gpd.GeoDataFrame(
usa_tracts,
columns=[self.TARGET_SCORE_RENAME_TO, "geometry"],
crs="EPSG:4326",
)
logger.info("Creating buckets from tracts")
usa_bucketed = self._create_buckets_from_tracts(
usa_tracts, self.NUMBER_OF_BUCKETS
)
logger.info("Aggregating buckets")
usa_aggregated = self._aggregate_buckets(usa_bucketed, agg_func="mean")
compressed = self._breakup_multipolygons(usa_aggregated, self.NUMBER_OF_BUCKETS)
self.geojson_score_usa_low = gpd.GeoDataFrame(
compressed,
columns=[self.TARGET_SCORE_RENAME_TO, "geometry"],
crs="EPSG:4326",
)
def _aggregate_to_tracts(
self, block_group_df: gpd.GeoDataFrame
) -> gpd.GeoDataFrame:
# The tract identifier is the first 11 digits of the GEOID
block_group_df["tract"] = block_group_df.apply(
lambda row: row["GEOID10"][0:11], axis=1
)
state_tracts = block_group_df.dissolve(by="tract", aggfunc="mean")
return state_tracts
def _create_buckets_from_tracts(
self, state_tracts: gpd.GeoDataFrame, num_buckets: int
) -> gpd.GeoDataFrame:
# assign tracts to buckets by D_SCORE
state_tracts.sort_values(self.TARGET_SCORE_RENAME_TO, inplace=True)
SCORE_bucket = []
bucket_size = math.ceil(len(state_tracts.index) / self.NUMBER_OF_BUCKETS)
for i in range(len(state_tracts.index)):
SCORE_bucket.extend([math.floor(i / bucket_size)])
state_tracts[f"{self.TARGET_SCORE_RENAME_TO}_bucket"] = SCORE_bucket
return state_tracts
def _aggregate_buckets(self, state_tracts: gpd.GeoDataFrame, agg_func: str):
# dissolve tracts by bucket
state_attr = state_tracts[
[
self.TARGET_SCORE_RENAME_TO,
f"{self.TARGET_SCORE_RENAME_TO}_bucket",
"geometry",
]
].reset_index(drop=True)
state_dissolve = state_attr.dissolve(
by=f"{self.TARGET_SCORE_RENAME_TO}_bucket", aggfunc=agg_func
)
return state_dissolve
def _breakup_multipolygons(
self, state_bucketed_df: gpd.GeoDataFrame, num_buckets: int
) -> gpd.GeoDataFrame:
compressed = []
for i in range(num_buckets):
for j in range(len(state_bucketed_df["geometry"][i].geoms)):
compressed.append(
[
state_bucketed_df[self.TARGET_SCORE_RENAME_TO][i],
state_bucketed_df["geometry"][i].geoms[j],
]
)
return compressed
def load(self) -> None:
logger.info("Writing usa-high (~9 minutes)")
self.geojson_score_usa_high.to_file(self.SCORE_HIGH_GEOJSON, driver="GeoJSON")
logger.info("Completed writing usa-high")
logger.info("Writing usa-low (~9 minutes)")
self.geojson_score_usa_low.to_file(self.SCORE_LOW_GEOJSON, driver="GeoJSON")
logger.info("Completed writing usa-low")

135
data/data-pipeline/data_pipeline/etl/score/etl_score_post.py Normal file
View file

@ -0,0 +1,135 @@
import pandas as pd
from data_pipeline.etl.base import ExtractTransformLoad
from data_pipeline.utils import get_module_logger
logger = get_module_logger(__name__)
class PostScoreETL(ExtractTransformLoad):
"""
A class used to instantiate an ETL object to retrieve and process data from
datasets.
"""
def __init__(self):
self.CENSUS_COUNTIES_ZIP_URL = "https://www2.census.gov/geo/docs/maps-data/data/gazetteer/Gaz_counties_national.zip"
self.CENSUS_COUNTIES_TXT = self.TMP_PATH / "Gaz_counties_national.txt"
self.CENSUS_COUNTIES_COLS = ["USPS", "GEOID", "NAME"]
self.CENSUS_USA_CSV = self.DATA_PATH / "census" / "csv" / "us.csv"
self.SCORE_CSV_PATH = self.DATA_PATH / "score" / "csv"
self.STATE_CSV = self.DATA_PATH / "census" / "csv" / "fips_states_2010.csv"
self.FULL_SCORE_CSV = self.SCORE_CSV_PATH / "full" / "usa.csv"
self.TILR_SCORE_CSV = self.SCORE_CSV_PATH / "tile" / "usa.csv"
self.TILES_SCORE_COLUMNS = [
"GEOID10",
"Score E (percentile)",
"Score E (top 25th percentile)",
"GEOID",
"State Abbreviation",
"County Name",
]
self.TILES_SCORE_CSV_PATH = self.SCORE_CSV_PATH / "tiles"
self.TILES_SCORE_CSV = self.TILES_SCORE_CSV_PATH / "usa.csv"
self.counties_df: pd.DataFrame
self.states_df: pd.DataFrame
self.score_df: pd.DataFrame
self.score_county_state_merged: pd.DataFrame
self.score_for_tiles: pd.DataFrame
def extract(self) -> None:
super().extract(
self.CENSUS_COUNTIES_ZIP_URL, self.TMP_PATH,
)
logger.info("Reading Counties CSV")
self.counties_df = pd.read_csv(
self.CENSUS_COUNTIES_TXT,
sep="\t",
dtype={"GEOID": "string", "USPS": "string"},
low_memory=False,
encoding="latin-1",
)
logger.info("Reading States CSV")
self.states_df = pd.read_csv(
self.STATE_CSV, dtype={"fips": "string", "state_code": "string"}
)
self.score_df = pd.read_csv(self.FULL_SCORE_CSV, dtype={"GEOID10": "string"})
def transform(self) -> None:
logger.info("Transforming data sources for Score + County CSV")
# rename some of the columns to prepare for merge
self.counties_df = self.counties_df[["USPS", "GEOID", "NAME"]]
self.counties_df.rename(
columns={"USPS": "State Abbreviation", "NAME": "County Name"}, inplace=True,
)
# remove unnecessary columns
self.states_df.rename(
columns={
"fips": "State Code",
"state_name": "State Name",
"state_abbreviation": "State Abbreviation",
},
inplace=True,
)
self.states_df.drop(["region", "division"], axis=1, inplace=True)
# add the tract level column
self.score_df["GEOID"] = self.score_df.GEOID10.str[:5]
# merge state with counties
county_state_merged = self.counties_df.merge(
self.states_df, on="State Abbreviation", how="left"
)
# merge state + county with score
self.score_county_state_merged = self.score_df.merge(
county_state_merged, on="GEOID", how="left"
)
# check if there are census cbgs without score
logger.info("Removing CBG rows without score")
## load cbgs
cbg_usa_df = pd.read_csv(
self.CENSUS_USA_CSV,
names=["GEOID10"],
dtype={"GEOID10": "string"},
low_memory=False,
header=None,
)
# merge census cbgs with score
merged_df = cbg_usa_df.merge(
self.score_county_state_merged, on="GEOID10", how="left"
)
# list the null score cbgs
null_cbg_df = merged_df[merged_df["Score E (percentile)"].isnull()]
# subsctract data sets
removed_df = pd.concat([merged_df, null_cbg_df, null_cbg_df]).drop_duplicates(
keep=False
)
# set the score to the new df
self.score_county_state_merged = removed_df
def load(self) -> None:
logger.info("Saving Full Score CSV with County Information")
self.SCORE_CSV_PATH.mkdir(parents=True, exist_ok=True)
self.score_county_state_merged.to_csv(self.FULL_SCORE_CSV, index=False)
logger.info("Saving Tile Score CSV")
# TODO: check which are the columns we'll use
# Related to: https://github.com/usds/justice40-tool/issues/302
score_tiles = self.score_county_state_merged[self.TILES_SCORE_COLUMNS]
self.TILES_SCORE_CSV_PATH.mkdir(parents=True, exist_ok=True)
score_tiles.to_csv(self.TILES_SCORE_CSV, index=False)