090°Open data
Open datasets, fully documented — searchable here, and readable by any LLM.
1 datasets
US Business Formation (Census BFS weekly state panel, keyless)
Weekly US business-formation panel, 2006 week 1 onward: the U.S. Census Bureau's Business Formation Statistics weekly state applications feed (keyless CSV; 51 entities x ~1,080 weeks), not seasonally adjusted. Complements the catalog's business_formation_monthly (monthly seasonally-adjusted, trailing 36-month window): this is the WEEKLY NSA grain with the full 2006-to-present history, for week-level joins and long training windows. Raw series per state-week: business applications (BA), high-propensity (HBA), with-planned-wages (WBA), from-corporations (CBA), plus the Census-published year-over-year percent changes. Deterministic derived features: HBA/WBA/CBA composition shares, national share of weekly applications, trailing 13-week and 52-week means, momentum_z (13-week mean vs trailing-year mean, standardized), dynamism_score 0-100 (cross-sectional percentile composite of momentum_z and YoY change; 50 = median state), formation_regime (surge/expansion/contraction/slump), signed streak_weeks, and trailing-52-week new high/low flags. Structural gates (exact layout, 51-entity balanced panel, contiguous weekly grids, non-negative counts, HBA/WBA/CBA <= BA, dates never in the future, freshness) fail loudly, never imputed. Census methodology note: as of the January 2026 release, internet-sales applications are excluded from the HBA/CBA series for the full time series. Who joins this: sales teams join week_end + state to time outreach into formation hotbeds; online shops read dynamism_score as a local-demand pulse; subscription businesses use momentum_z and hba_share as SMB-formation features for churn and expansion models.
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.