Problem
The canonical congressional-district GeoJSON lives at
policyengine-app-v2/app/public/data/geojson/congressional_districts.geojson
— a frontend repo. Every PolicyEngine dashboard that renders a
choropleth (NC-Stein, WPTRA, NJ CTC/EITC, WV SB392, RCC, etc.) bundles
a copy. The per-district h5 files live here, but the polygons used to
build them live elsewhere.
This pretty much guarantees drift between:
- the polygons used to sample households into districts (the h5 build),
- the polygons rendered as the choropleth in each dashboard, and
- the polygons that hardcoded representative / region maps were
authored against.
Ask
Why now
Companion to #1143 (boundary-vintage metadata) and #1144 (119th → 120th
transition). Both are much harder to coordinate if the geojson is in a
different repo with a different release cadence than the data it
represents.
Related
Problem
The canonical congressional-district GeoJSON lives at
policyengine-app-v2/app/public/data/geojson/congressional_districts.geojson— a frontend repo. Every PolicyEngine dashboard that renders a
choropleth (NC-Stein, WPTRA, NJ CTC/EITC, WV SB392, RCC, etc.) bundles
a copy. The per-district h5 files live here, but the polygons used to
build them live elsewhere.
This pretty much guarantees drift between:
authored against.
Ask
congressional_districts.geojsonintopolicyengine-us-dataalongside the h5 build script that consumes it.
set (e.g.,
policyengine/policyengine-us-data/districts/geojson/{congress}.geojson).flag the local copy as a cache.
geojson from a single
@policyengine/district-boundariespackage orHF URL.
Why now
Companion to #1143 (boundary-vintage metadata) and #1144 (119th → 120th
transition). Both are much harder to coordinate if the geojson is in a
different repo with a different release cadence than the data it
represents.
Related