For civic tech & journalism

Census, election, and open-data geography for newsrooms

All the geography a journalist, researcher, or non-profit needs — states, counties, congressional districts (113th–119th Congresses), state legislative districts (upper and lower chambers), Census tracts (~84,000), ZIP code areas (~33,000), unified school districts (~10,000) — joined to Census demographics and your own datasets. Voting precincts available via one-call upload from your workspace. Free for 50 calls per month, forever.

What it looks like

A typical workflow for civic tech & journalism teams. Each line is one operation; together they take a few minutes and produce a published map plus a downloadable dataset.

# A data reporter on deadline —
# "Make me a county-level map of voter turnout vs. median age."

mcp.call("list_shape_versions", { collection: "us-counties" })
  → tiger-2024 (3,143 counties)

mcp.call("census_acs", {
  collection: "us-counties", version: "tiger-2024",
  year: 2023, variables: ["B01002_001E"],   // median age
  slug: "counties-median-age-2023",
})
  → median age per county (workspace dataset)

mcp.call("ingest_county_pres", { year: 2024 })
  → workspace dataset with turnout per county

mcp.call("create_report", {
  slug: "turnout-vs-age", name: "Turnout × median age (counties, 2024)",
  public: true,
  config: {
    collection: "us-counties", version: "tiger-2024",
    datasets: [{ slug: "counties-median-age-2023", field: "B01002_001E" }],
    render: "bivariate",   // see render_map_bivariate
  }
})
  → /v/turnout-vs-age   embeddable in newsroom CMS

mcp.call("describe_report", { slug: "turnout-vs-age" })
  → "Higher-median-age counties (top quintile) had turnout
     of 71%, vs. 58% in the youngest quintile..."

What you can build

The tools you'd actually use

list_shape_collections
Browse the catalog — states, counties, congressional districts, tracts, precincts
ingest_county_pres
Pre-built 2024 / 2020 / 2016 county presidential results
census_acs
Any Census American Community Survey table for any geography level (state / county / congressional district / tract / ZIP code / place)
disaggregate_acs
Push Census tract-level data down to precincts or any custom polygon — area-weighted interpolation in one call
create_report
Save a viewer config, publish at /v/<slug>
render_map_bivariate
Two-variable map for nuanced storytelling
build_geojson
Bulk map-file export for desktop mapping software
describe_report
Claude writes a publication-ready paragraph about your map
compare_versions_narrative
Generate the explainer for any redistricting change

How it differs from the alternative

vs. The Big Local News / Sunlight / Census Reporter: focused on map-making, not data discovery. vs. building from raw Census boundary files: every join is one call, not a notebook. Free tier is enough to make a real election-night map; no upsell coercion.

Pricing for civic tech & journalism teams

Most newsrooms run on Free (50 calls per month is enough for a few real maps). Heavier research projects fit on Starter ($10/month). Full pricing details.

Free tier, real geography.

Build a real map without leaving Claude.

Get an API key

The same toolkit also covers grassroots organizing platforms & campaign tooling · healthcare & life sciences · real estate & property tech · retail, consumer brands & site selection · insurance & risk modeling · logistics, delivery & field service · research, academia & think tanks · pollsters & survey research firms — one U.S.-geography surface for any team whose question comes down to where.