Name: README.md
Size: 7.52 KB
Content Type: text/markdown; charset=utf-8
Last Modified: 16 Jun 2026
Source URL: https://data.source.coop/cboettig/ca-dac/README.md
Cloud URI: s3://us-west-2.opendata.source.coop/cboettig/ca-dac/README.md

California Disadvantaged Communities (DAC) and Economically Distressed Areas (EDA), 2023

This dataset republishes California Department of Water Resources (DWR) layers for Disadvantaged Communities (DAC) and Economically Distressed Areas (EDA) at three Census geographies (Place, Tract, Block Group), with County also available for EDA. Source data are joined from American Community Survey (ACS) 2019–2023 5-year estimates onto 2020 Census geographies, and identify communities eligible for California water-related funding programs.

A community is designated Disadvantaged (DAC) when its Median Household Income (MHI) is less than 80% of the statewide MHI per California Water Code §79505.5(a). The published DAC23 field encodes this as Y / N / Data Not Available. The finer Severely Disadvantaged (SDAC) distinction (MHI < 60% of statewide) is not encoded as a separate flag in this dataset — derive it from MHI23 and the statewide MHI threshold relevant to your eligibility window.

Economically Distressed Areas (EDA) are defined per Water Code §79702(k) as a subset of municipalities with MHI < 85% of statewide MHI together with at least two of: financial hardship, unemployment ≥ 2% above statewide, or low population density.

Layers

Files are organized by the published Census geography:

Layer	Path prefix	Features	Use
DAC — Census Tract	`dac-2023/tract`	9,129	Most common eligibility unit
DAC — Census Place	`dac-2023/place`	1,618	Cities and CDPs
DAC — Block Group	`dac-2023/blockgroup`	25,607	Finest-grain DAC layer
EDA — Census Tract	`eda-2023/tract`	9,129	EDA at tract level
EDA — Census Place	`eda-2023/place`	1,618	EDA at place level
EDA — Block Group	`eda-2023/blockgroup`	25,607	EDA at block-group level
EDA — County	`eda-2023/county`	58	EDA at county level

Each layer is published in three formats:

<prefix>.parquet — GeoParquet for analytical queries (DuckDB, Polars, geopandas)
<prefix>.pmtiles — Vector tiles for web maps (MapLibre, Leaflet)
<prefix>/hex/h0={cell}/data_0.parquet — H3 hexagonal tiling at resolution 10, hive-partitioned by H3 resolution-0 parent cell

Field Reference

DAC fields (Tract, Place, Block Group)

Field	Type	Description
`GEOID` / `GEOID20`	string	Census FIPS code (state+county+tract+blkgrp, or place code)
`STATEFP` / `STATEFP20`	string	State FIPS (`06` for California)
`COUNTYFP20`	string	County FIPS (Tract, Block Group only)
`TRACTCE20`	string	Tract code within county
`BLKGRPCE20`	string	Block group code within tract (Block Group only)
`PlaceCode`	string	Place code (Place only)
`PlaceNames`	string	Comma-separated place names (Place only)
`ALAND` / `ALAND20`	float	Land area in square meters
`AWATER` / `AWATER20`	float	Water area in square meters
`INTPTLAT` / `INTPTLON`	string	Internal point latitude / longitude
`Pop23`	float	Total population, ACS 2019–2023
`HH23`	float	Households, ACS 2019–2023
`MHI23`	float	Median household income, ACS 2019–2023 (USD)
`DAC23`	string	DAC flag: `Y` (designated DAC), `N` (not DAC), or `Data Not Available`. The Tract layer also distinguishes severely-disadvantaged designations — inspect distinct values in your layer of interest.
`Source`	string	Data source for the MHI estimate
`Comments`	string	Editor notes (often empty)
`Date_Recor`	date	When the row was last edited (truncated from shapefile DBF)
`Record_Las`	string	Editor identifier (truncated from shapefile DBF)
`Date_Data_`	string	Vintage of the underlying ACS estimate (truncated from shapefile DBF)
`_cng_fid`, `OGC_FID`	int64	Row-unique keys added by the conversion pipeline

EDA-only fields (in addition to the geographic identifiers above)

Field	Type	Description
`LandSqMile`	float	Land area in square miles
`PopDensity`	int	Population density (people per land sq mile)
`Flag_MHI_P`	string	Flag indicating MHI is below 85% of statewide MHI
`Unempl_EDD`	float	Unemployment rate (Place, County)
`UnEmFlgEDD`	string	Flag indicating unemployment ≥ 2 points above statewide (Place, County)
`NAMELSAD` / `NAMELSAD20`	string	Geography name (Place, County)

Note: Source data are published as shapefiles, where DBF column names are capped at 10 characters. Names longer than 10 characters in the original DWR schema are truncated in the GeoParquet (e.g. Date_Record_Last_Edited → Date_Recor, Shape.STArea() → ShapeSTAre).

H3 indexing fields (added by this pipeline)

Field	Type	Description
`h10`	uint64	H3 cell at resolution 10 (~15,000 m² per hex)
`h9`, `h8`	uint64	Parent H3 cells
`h0`	uint64	H3 resolution-0 parent (partition key)

Usage

DuckDB — analytical query

DuckDB — H3 hex spatial join

MapLibre — DAC tracts overlay

PMTiles internal source-layer names by file:

File	`source-layer`
`dac-2023/tract.pmtiles`	`tract`
`dac-2023/place.pmtiles`	`place`
`dac-2023/blockgroup.pmtiles`	`blockgroup`
`eda-2023/tract.pmtiles`	`tract`
`eda-2023/place.pmtiles`	`place`
`eda-2023/blockgroup.pmtiles`	`blockgroup`
`eda-2023/county.pmtiles`	`county`

Citation

California Department of Water Resources (DWR). Disadvantaged Communities and Economically Distressed Areas Mapping Tool, 2023 update. https://gis.water.ca.gov/app/dacs/

Original data layers are republished from the California Natural Resources Agency Open Data hub under the State of California's terms of use.

Data Lineage

Source: DWR ArcGIS hosted layers, 2023 vintage, downloaded as shapefile (EPSG:4326) from the CNRA Open Data hub
Conversion: cng-convert-to-parquet produces optimized GeoParquet
PMTiles: tippecanoe with --drop-densest-as-needed --extend-zooms-if-still-dropping
H3 hex tiling: cng-datasets vector at resolution 10, parent resolutions 9, 8, 0, hive-partitioned on h0

See boettiger-lab/datasets for the processing pipeline source.