Utah Healthcare Workforce Intelligence Dashboard

01 / Problem

Utah healthcare organizations making hiring decisions are working from national labor statistics or raw FRED series that require significant preprocessing to be useful. There was no decision-ready, Utah-specific view that combined macro leading indicators with local employment data and a forward-looking forecast.

02 / Approach

The backend is FastAPI pulling from the FRED API — national JOLTS series, wage and CPI data, and Utah-specific healthcare employment figures. A YAML-configured preprocessing pipeline handles frequency alignment and differencing. SARIMAX produces 6-month employment and quit-rate forecasts, and a Composite Stress Score (CSS) synthesizes the leading indicators into a single Tightening / Stable / Easing signal. SQLite stores series with revision tracking; a scheduler handles daily data refreshes and monthly model retraining. The React + Vite frontend has six sections: the CSS headline, JOLTS sparklines, wage pressure charts, an employment trend forecast, a 6-row forecast table, and a methodology transparency panel.

03 / Result

The dashboard gives hiring decision-makers a single page that goes from raw macro data to an actionable signal with a 6-month forecast. The backend serves realistic dummy data by default, so the full frontend works immediately without a FRED key or trained model — useful for demos and local development.