Explainable mood inference from physiological data
Now live at: https://swacziarg.github.io/HappyApp/#/login
HappyApp is a secure, multi-user web app that converts raw Garmin health data into personalized, explainable daily mood predictions.
- Users upload Garmin exports (ZIP / JSON)
- Backend ingests and normalizes physiological data
- Personalized baselines are computed per user
- Daily features are inferred into a mood score (1–5)
- Users can add manual mood check-ins
- A React UI displays today’s mood and historical trends
Garmin Export → Ingestion → Daily Features (personal baselines) → Mood Inference (rule-based) → Predictions API → React UI
Backend: FastAPI, Supabase (Postgres + RLS)
Frontend: React
Auth: Supabase Auth (JWT)
Deployment: Render + GitHub
Data: Garmin exports (sleep, HRV, stress, activity, body battery)