An AI-assisted, context-aware health awareness platform integrating environmental data, behavioral reflection, and Retrieval-Augmented Generation (RAG) to promote sustainable well-being, aligned with UN SDG 3.
Maintaining healthy lifestyle habits in modern urban environments is increasingly challenging due to academic and professional stress, environmental degradation, and fragmented access to reliable health information. This project presents a modular, AI-assisted web platform that enables preventive health awareness by combining user-reported lifestyle data, real-time environmental context, and ethically constrained AI-driven explanations.
The system is explicitly designed as a decision-support and awareness tool, not a diagnostic or clinical system, adhering strictly to Responsible AI principles.
-
Primary SDG: SDG 3 – Good Health and Well-being
-
Secondary Alignment: SDG 11 – Sustainable Cities and Communities (via air quality and environmental health awareness)
🌐 Live Application:
👉 https://sdg3-health-awareness-app.onrender.com/
Despite the availability of health-related data and digital tools, individuals often lack integrated, interpretable, and context-aware insights that connect daily habits with environmental conditions. Existing solutions frequently focus on medical outcomes rather than preventive awareness, leading to low engagement and ethical concerns.
How might we leverage Responsible AI to synthesize lifestyle and environmental signals into accessible health awareness insights that support sustainable daily decision-making without replacing professional healthcare?
The platform is implemented as a full-stack Flask application following a service-oriented architecture, enabling clean separation between user interaction, business logic, AI services, and knowledge retrieval.
- Lifestyle data acquisition via structured daily check-ins
- Environmental context integration (weather & air quality)
- Seasonal and behavioral awareness modeling
- AI-powered conversational interface with RAG grounding
- Ethical safeguards and transparency mechanisms
AI is employed strategically and conservatively to enhance interpretability and scalability rather than automation of medical decisions.
-
Conversational AI Interface Uses prompt-engineered LLM interactions to answer general health-related queries.
-
Retrieval-Augmented Generation (RAG) AI responses are grounded in a curated local health knowledge base, ensuring:
- Reduced hallucination risk
- Improved factual consistency
- Transparent information sourcing
-
Explainability-First Design AI outputs are contextualized, non-prescriptive, and aligned with preventive health education.
⚠️ The system does not perform diagnosis, prediction of disease, or treatment recommendation.
| Layer | Technology |
|---|---|
| Backend | Flask (Python) |
| Frontend | HTML5, CSS3 |
| AI Layer | OpenRouter (LLM API) |
| Knowledge Grounding | Local RAG (Text-based Retrieval) |
| External Data | Weather & Air Quality APIs |
| Architecture | Modular, Service-Oriented |
| Version Control | Git & GitHub |
healthy-lifestyle-awareness-ai/
├── app/
│ ├── routes/ # Request routing & controllers
│ ├── services/ # Business logic, AI, RAG modules
│ ├── templates/ # Jinja2 UI templates
│ ├── static/ # CSS, images, UI assets
│ └── knowledge/ # Curated health knowledge base
├── run.py # Application entry point
├── requirements.txt
├── .gitignore
└── README.md
This project was explicitly designed to align with Responsible AI guidelines:
- Fairness: No personalized medical conclusions or demographic assumptions
- Transparency: Clear disclosure of AI-generated content and system limitations
- Ethical Use: AI restricted to awareness, education, and explanation
- Privacy Preservation: No storage of personal or sensitive user data
- Risk Mitigation: RAG-based grounding to reduce misinformation and hallucinations
- Increased preventive health awareness
- Improved understanding of environmental health risks
- Reduction in misinformation through grounded AI explanations
- Support for mental well-being via habit reflection
- Promotion of sustainable lifestyle behaviors
- Students and young professionals
- Urban populations exposed to environmental stressors
- Communities seeking accessible health awareness tools
AI services are configured using environment variables to prevent credential leakage:
OPENROUTER_API_KEY=your_api_key_hereSecrets are intentionally excluded from version control.
python -m venv venv
source venv/bin/activate # Linux/Mac
venv\Scripts\activate # Windows
pip install -r requirements.txt
python run.pyThis project demonstrates that effective AI for sustainability does not require complex models, but rather:
- Clear problem framing
- Ethical constraints
- Context-aware system design
- Explainability and transparency
It reinforces AI as a supportive socio-technical tool, not a replacement for human judgment.
Developed as part of the 1M1B – IBM SkillsBuild AI for Sustainability Virtual Internship
Dipanshu R. Shamkuwar
B.Tech Computer Science & Engineering