Atlas · Economic impact
A small intervention — a broadband line, a clinic, a $50K grant — propagates through county-level employment, wage, and outcome data. Modeled across 3,144 US counties. Monte Carlo bands published.
The Atlas Butterfly Effect engine answers a specific question: given an intervention X (a broadband investment, a clinic, a credit program, a curriculum) at location Y at time T, what is the probability distribution of downstream economic outcomes Z over years t+1 through t+10?
It does this by composing four signal channels at the county FIPS level: broadband infrastructure (3,144 counties), employment + wages (BLS OEWS, 82,522 SOC × NAICS rows), education program completion (CIP scorecard, 14,259 program-credentials), and capital flow projections (3 Monte Carlo runs).
Below is the most recent published Monte Carlo run from atlas_capital_flow_projections. The 24.1% feudalism probability is the share of simulation paths where capital concentration exceeds the published threshold; the 99.75% inversion probability is the share of paths where gini reverses by year t+10.
Conventional impact analysis publishes a single point estimate ("this program created N jobs"). The Atlas instead publishes a probability distribution and the channel decomposition, so a reader can see exactly which assumptions are doing the load-bearing work. We rate methodology, not conclusions.
An honest economic claim is shaped like a distribution, not a number. Anyone publishing only a number is asking you to trust the median quietly.
The county-level data is queryable via atlas_broadband_access, atlas_oews, and atlas_capital_flow_projections directly, or through the operational Butterfly tool on DaedArch for any FIPS code.