Inspiration

What it does## Inspiration

Valentine’s Day is supposed to be “happy,” but a lot of people actually feel extra pressure, loneliness, or emotional swings around it. We wanted to build something that feels gentle, quick, and non-judgmental—a tool that helps you notice patterns in your emotional health and take small steps before stress quietly builds up.

What we built

LoveCare is a lightweight emotional health check-in app. In under a minute, users log:

  • mood, stress, energy, sleep

The app turns these signals into a simple weekly snapshot and a clear, supportive summary, including:

  • trends (e.g., mood downtrend, sleep deficit)
  • volatility (how much things swing day-to-day)
  • a risk label (Green / Yellow / Red) with reasons and micro-actions
    > Disclaimer: LoveCare provides wellness insights only — not medical diagnosis.

How it works (high level)

We designed a simple scoring system that balances interpretability and usefulness.

  • Baseline levels: average mood/stress/energy and sleep compared to a target (e.g., 8 hours).
  • Trends: we check whether mood/energy is steadily dropping over the week.
  • Volatility: we measure how unstable the daily signals are (large swings can indicate strain even if the average looks “fine”).
  • Risk score → label: we map the combined score to Green / Yellow / Red and generate human-readable reasons.

This keeps the logic transparent so users can understand why a label was assigned.

Tech stack

  • Frontend: React (TypeScript)
  • Backend: FastAPI (Python)
  • Storage: lightweight database (SQLite) for users and check-ins
  • API: endpoints for login, bulk check-ins, and report generation

Challenges

  • Designing “helpful but safe” insights: we avoided medical claims, focused on supportive language, and always included a disclaimer.
  • Scoring calibration: we iterated on weights so that one bad day doesn’t overreact, but consistent sleep deficit/high stress does surface.
  • Clean UX under hackathon time: keeping the flow fast (check-in in < 60 seconds) while still producing meaningful output.

What we learned

  • Simple, interpretable metrics can feel more trustworthy than a black-box model for wellness use cases.
  • Good product language matters: tone + framing can change whether users feel supported or judged.
  • Building end-to-end (frontend + backend + scoring) taught us how to ship a cohesive experience quickly.

Next steps

  • Personalization (user-set sleep targets, different baselines)
  • Optional reminders and streaks (without guilt)
  • Better reflection analysis (theme extraction, mood triggers)
  • More accessibility features and privacy controls

Built With

Share this project:

Updates