Source code at Github

Personal Well-Being & Lifestyle Intelligence Platform

A privacy-first, non-clinical, AI-powered lifestyle and well-being system

This project aims to build a Personal Well-Being & Lifestyle Intelligence Platform that helps users (initially students) understand their habits, energy patterns, and stress signals early — without diagnosing or medicalizing mental health.

The system fuses objective behavioral data (sleep, screen time, activity) with subjective inputs (mood, journaling) to form a digital phenotype of the user’s lifestyle state. Machine learning is used only for pattern detection, risk trends, and personalized nudges, not diagnosis or treatment.

Vision

Modern burnout and chronic stress are not sudden events — they emerge gradually from lifestyle patterns such as poor sleep consistency, excessive screen use, cognitive overload, and lack of recovery.

This platform is designed to:

• Detect early warning signals of overload
• Help users understand how they live, not label them
• Provide data-backed, personalized lifestyle nudges
• Preserve privacy, autonomy, and trust

Positioning:

Lifestyle intelligence tool, not a medical or therapeutic system

Core Principles

• Non-clinical: No diagnosis, no therapy language
• Privacy-first: Local-first analytics, federated learning
• User control: Modular features, pause/opt-out anytime
• Interpretability: Explain why a signal appears
• Long-term value: Habit intelligence over short-term alerts

High-Level Capabilities

1. Digital Phenotyping (Behavior + Self-Report)

The system builds a continuous digital phenotype using:

Objective Signals

• Sleep duration & interruptions (HealthKit / Google Fit)
• Screen time & app category usage
• Physical activity (steps, workouts)
• Daily workload indicators (manual + calendar)

Subjective Signals

• Daily mood check-ins
• Optional journaling
• Self-reported workload & stress perception

Together, these form a multi-dimensional lifestyle state instead of isolated metrics.