Rendered with OpenLayers over our own PostGIS-served boundaries and vector tiles — GeoJSON and MVT straight from the gold warehouse, no third-party basemap. Partisan choropleths are drawn equal-area (EPSG:5070) so shaded area is honest.
Drill down through states → congressional districts → counties → precincts on one map, then open any
geography's HUD: demographics, elections and money-in at a glance, with its candidate
roster linked out. Every geography is a page (/explore/geo/<level>/<id>/) — e.g.
geo/state/CA, geo/cd/CA-27,
geo/county/06037.
Build a question over places, then read it as a table, a ranking, a counter or a map — the same dataset, a curated set of views. Add a measure to make a column, add a clause to filter. Partisan & registration lean, turnout, CVAP and population composition, at state or county grain. Shareable spec.
newThe analytical surface: colour every county by any variable (partisan lean, votes, population, CVAP composition) with your own classification — natural breaks, quantiles, equal interval or manual — then add filters on any other variables to isolate a subset. The first slice of the full data catalog; more sources & vintages to come.
newPut any counties side by side across every variable at once — the spread in each row is highlighted. Shareable. Comparing a place to its own past arrives with historical vintages.
liveEvery county shaded Dem↔Rep by (Dem−Rep)÷total, aggregated up from precinct returns. Equal-area.
liveCongressional districts and state-legislative chambers, served as vector tiles. Switch layers in-map.
live175k RDH precincts shaded by (Dem−Rep)÷total, per state. Lean year shown per precinct (2020–2024, each state's latest-loaded general) — honest, not silently mixed.
live85k census tracts, ACS estimate. Switch the measure AND classify it yourself — natural breaks, quantiles, equal interval or manual bands, your class count and colour ramp. QGIS-style, in the browser.
liveA bivariate choropleth over a 3×3 key: pick ANY two variables (lean, votes, population, CVAP composition) for the X and Y axes and see both at once. Defaults to diversity × 2020 lean — the interesting corners pop.