💡 Inspiration
Early warning signs in neurological health are easy to miss, and updates often get lost between patients, families, and clinicians.
We set out to make early signals visible (detect), communication structured (connect), and suggestions personalized in minutes.
⚙️ What it does
- Capture: Log sleep, mood, headache, memory lapses, stress, and exercise in seconds.
- Detect: A transparent 0–100 risk score highlights early patterns and trends.
- Connect: Generates a concise 14-day doctor summary + CSV export for clinical review.
- Personalize: Provides adaptive tips and insights from each entry and shows trajectory.
🛠️ How we built it
A fully client-side web app using HTML, CSS, and vanilla JavaScript.
Data is stored locally in the browser (localStorage).
Implemented a lightweight canvas-based chart and an interpretable rule-based risk model with a trend component.
No external dependencies or backend required.
📚 What we learned
- UX makes or breaks health logging — speed and clarity drive adoption.
- Simple, interpretable signals become useful when combined with trend context.
- Local-first prototypes reduce friction and improve privacy in a hackathon setting.
🚧 Challenges
- Shipping a polished prototype under severe time constraints.
- Balancing interpretability and usefulness (kept the scoring simple and documented).
- Designing a summary that is short enough for clinicians yet information-dense.
🔮 What’s next
- Optional secure cloud sync + team/caregiver sharing.
- Clinician dashboard, HL7/FHIR export, and data validation.
- Optional LLM layer for longer-term summaries and anomaly explanations.
- Formal usability testing and accessibility review.
🛡️ Safety & Ethics
- Non-diagnostic: outputs are heuristic, not a medical diagnosis.
- Transparency: risk model weights are fully documented.
- Guidance: high-risk results include clear advice to seek professional care.
- Privacy: local-first by default; no backend data collection.
Log in or sign up for Devpost to join the conversation.