About
This site lets you ask free-text questions about CDC mortality and cancer data. The backend is a FastAPI service that uses an OpenAI model to translate the question into SQL, runs it against ~4.95M rows of CDC data in PostgreSQL, then drafts a written answer plus an optional Vega-Lite chart spec.
Data sources: USCS (US Cancer Statistics) bulk files for cancer, plus several NCHS Socrata datasets for mortality, life expectancy, weekly state deaths, and excess deaths.
Coverage caveats: pre-1999 mortality is limited; some recent years (esp. 2023) are partial in the weekly excess-deaths dataset; county-level cancer is local-only.