Vernal.earth is a county-level environmental exposure atlas covering all 3,221 US counties, paired with an AI-powered research assistant for testing whether environmental conditions predict health and wildlife disease outcomes.
The platform combines data from federal agencies (USGS, EPA, CDC, Census, NASA, NOAA) into a unified database with a correlation engine, regression analysis, and spatial queries. All analyses include statistical rigor indicators: sample sizes, p-values, effect sizes, and Bonferroni correction for multiple comparisons.
Vernal does not prove that environmental contamination causes disease. County-level ecological studies identify geographic associations and generate hypotheses. Proving causation requires individual-level epidemiology, controlled experiments, or mechanistic validation at the molecular level.
What Vernal does is make hypothesis generation fast, reproducible, and grounded in real federal data — reducing the time from "I wonder if..." to "here's the statistical association, here's the mechanism, here's what to test next" from months to hours.
PostgreSQL + PostGIS, FastAPI, SvelteKit, Claude API. One person, AI tools, federal data, and a Mac Mini. April 2026.