Data & Download

Source Data

Where the Data Comes From

All HMDA data is sourced directly from the Consumer Financial Protection Bureau's public HMDA Data Browser API. No modeling, no imputation — this is directly reported lender data.

CFPB HMDA Data Browser

The Consumer Financial Protection Bureau collects and publishes loan-level HMDA records annually. The HMDA Data Browser provides a REST API for downloading filtered record sets by state, year, and institution. Plat downloads all home-purchase originations for all 50 states and the District of Columbia.

ffiec.cfpb.gov/data-browser →

Census TIGER/Line Shapefiles

Geographic boundaries for census tracts, counties, and states come from the U.S. Census Bureau's TIGER/Line program. Plat uses three boundary vintages (2000, 2010, 2020) aligned to the decennial census tract definitions in use for each year of HMDA data.

Census TIGER/Line Files →

Aggregated Files

Plat Data Files

Plat pre-aggregates raw HMDA records into compact JSON files — one file per geographic level per year. These are the same files the interactive maps load. You can use them directly in your own analysis.

Bulk download links are coming soon. In the meantime, you can access all JSON files directly from the data/ directory of the Plat repository.

states_{year}.json

State-level aggregates for a single year. 51 records (50 states + D.C.), <10 KB each.

18 files · ~150 KB total JSON
counties_{year}.json

County-level aggregates for a single year. ~3,200 county FIPS keys per file.

18 files · ~25 MB total JSON
tracts_{epoch}_{year}.json

Census tract aggregates for a single year. ~84,000 GEOID keys. Loaded on demand when map zooms to tract level.

18 files · ~5 MB each JSON
homebuyers.json

National totals by year and racial group — used for the loan-volume bar chart.

<10 KB JSON
income.json

National median income by year and racial group — used for the income scatter chart.

<10 KB JSON
states.pmtiles / counties.pmtiles

Vector tile archives for state and county boundaries. Streamed via HTTP range requests — no tile server required.

~5 MB combined PMTiles

Schema

Data Field Reference

Each record in a state, county, or tract JSON file contains the following fields. All income values are in nominal U.S. dollars (not inflation-adjusted).

Field	Type	Description
r	string	Dominant racial/ethnic group (Asian, Black, Hispanic, White, Multi, or Diverse)
tx	integer	Total home-purchase loan originations for this geography and year
inc	integer	Median applicant income across all originations (dollars)
pw	float	Share of originations where applicant race is White (0–1)
pb	float	Share of originations — Black or African American
pa	float	Share of originations — Asian
ph	float	Share of originations — Hispanic or Latino
pm	float	Share of originations — Multiracial or other
iw	integer	Median income — White applicants (dollars, null if <5 records)
ib	integer	Median income — Black applicants
ia	integer	Median income — Asian applicants
ih	integer	Median income — Hispanic applicants
im	integer	Median income — Multiracial applicants

Terms

Data Use & Attribution

Plat's aggregated data files are made available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You are free to use, share, and adapt them for any purpose, including commercial use, as long as you provide attribution.

The underlying HMDA raw data is a U.S. federal government work and is in the public domain. No license restrictions apply to the raw CFPB data itself.

Attribution

When using Plat data in published work, please include: "Data aggregated by Plat from CFPB HMDA public records." See the About page for a full citation format.

Limitations

HMDA data reflects applications made to covered lenders — not all mortgage activity, and not all housing transactions. See the Methodology page for a full description of known data limitations.