090°Open data
Open datasets, fully documented — searchable here, and readable by any LLM.
1 datasets
US County Poverty & Income Intelligence
Census Bureau SAIPE state-and-county bulk files (keyless official text downloads, U.S. public domain): 510 per-state files stacked into a county x estimate-year panel (3,140-3,142 publishable counties per vintage, 31,408 rows over 2015-2024) with all-ages and child (0-17) poverty counts and rates plus 90% confidence intervals, median household income with intervals, year-over-year and 5-year poverty-rate momentum in percentage points and income YoY percent change. A documented 0-100 household-stress score (45% all-ages poverty rate + 35% child poverty rate + 20% inverse median income, winsorized at 1/99 and min-maxed within each year) is ranked nationally (stress_rank) and within each state, tiered s1-s4, with persistent-high-poverty (>=20% for 3 straight years), improving (YoY <= -1pp) and worsening (YoY >= +1pp) flags. Primary key: (county_fips, estimate_year). Cadence: yearly; each estimate year comes from its own SAIPE release (vintage stack, methodology revisions not restated), as_of is the latest vintage's file-release date and identical input produces an identical content hash. Caveats: model-based estimates (use the CI columns for significance); county-boundary revisions (notably Connecticut 2022) yield null momentum, never fabricated links. Sample use: order by stress_rank within an estimate year for the most stressed counties, or filter state_code for within-state household-strain geography.
Once today's free allowance is used up, AI features can run on your own provider account.
Kept in this browser tab only (cleared when you close it) and sent with each AI request. Our servers use it for that request and never store or log it.