Time-series forecasting dashboard for monthly claims and paid amounts using SARIMAX models on OSHA severe-injury report data.
- Claims & Cost Forecasting – Monthly charts with confidence bands powered by SARIMAX seasonal models
- Segment Filtering – Filter by state, industry, claim type, date window, and forecast horizon
- Scenario Analysis – What-if recalculation with severity inflation and frequency shock adjustments
- Monthly Data Table – Tabular view aligned with forecast charts
- AI Summary – Natural language forecast interpretation via OpenAI (deterministic fallback if unavailable)
- Data Pipeline – OSHA Severe Injury Report ingestion, merge, and synthetic gap filling
The application applies SARIMAX (1,1,1)x(1,1,1,12) seasonal time-series models to OSHA-derived severe-injury report data. Users select segments (state, industry, claim type) and a forecast horizon, and the system generates point forecasts with 95% prediction intervals. A scenario panel allows adjusting severity inflation and frequency shocks to model what-if situations. If SARIMAX fails due to short or noisy series, the system falls back to ARIMA or seasonal naive methods.
| Category | Technologies |
|---|---|
| Backend | Python 3.13, FastAPI, Uvicorn |
| Frontend | TypeScript, Next.js, React, Tailwind CSS |
| AI/ML | SARIMAX, OpenAI |
| Data | pandas, NumPy, statsmodels |
| Charts | Recharts |
| Package Management | uv (backend), pnpm (frontend) |
| Deployment | Docker, GitHub Actions, Google Artifact Registry |
# Backend
cd backend
cp .env.example .env # Add your OpenAI API key
uv sync
uv run uvicorn insurecast.main:app --reload
# Frontend
cd frontend
pnpm install
pnpm devOpen http://localhost:3000 (frontend) and http://localhost:8000/docs (API docs).