A frontend prototype demonstrating how AI can transform verified lab data into accurate predictions, reduced manual effort, and smarter environmental decisions.
HMPI is a frontend-first prototype designed to showcase an AI-centric approach to solving real problems faced by researchers and laboratories:
- manual HMPI calculations
- repetitive report generation
- fragmented historical data
- lack of predictive insights
Our focus is not just calculation — it is understanding, prediction, and prevention.
Researchers currently:
- perform manual calculations
- face human errors
- lack historical & predictive context
- work with isolated datasets
This prototype demonstrates how an AI-driven system can:
- automate analysis
- reduce errors
- efficiently reuse historical data
- support future decision-making
⚠️ Note:
The AI model and backend pipelines described below are conceptual designs
intended to demonstrate feasibility and future scope.
They are not fully implemented in this prototype.
- 📈 Predict future HMPI / pollution trends
- 🚨 Detect anomalies in lab data
- 📊 Learn from historical, district-level datasets
- 🧩 Convert raw lab uploads into actionable insights
The AI model is the engine — everything else supports it.
- Researcher dashboard UI
- CSV upload interface (mocked flow)
- HMPI calculation visualization
- Graphs & analytical views
- Heatmap & map-based visualization
- File history & activity UI
- Clean, responsive design (SIH-ready)
- Backend APIs
- Database pipelines
- AI/ML models
- Lab verification APIs
- Large-scale indexing & optimization
-
MongoDB (NoSQL)
- Raw CSV uploads
- Flexible lab schemas
- High-volume ingestion
-
PostgreSQL + PostGIS
- Clean structured data
- Spatial queries (district-wise)
- Required for AI training
MongoDB stores what is uploaded
SQL stores what the AI learns from
- CSV files contain only latitude & longitude
- Reverse geocoding APIs determine district automatically
- Enables:
- district-level analysis
- historical comparisons
- map-based intelligence
Planned APIs
- OpenStreetMap Nominatim (free)
- Geoapify / LocationIQ (optional)
Planned verification via:
- NABL
- CPCB
- SPCB
Goal:
- prevent fake lab uploads
- ensure data credibility
- build national-level trust
- 📂 CSV upload flow (UI)
- 🧮 Automated HMPI visualization
- 🗺️ Heatmaps & spatial analysis
- 📉 Trend graphs
- 📄 Report preview UI
- 🧠 AI prediction (conceptual)
- 🕒 Upload history & activity tracking
- Indexed district-level searches
- TB-scale data handling (conceptual)
- Fast historical lookups using indexing
- Optimized for government-scale datasets
Designed today to scale tomorrow.
- Code versioning via GitHub
- AI model versioning (rollback support)
- Dataset version tracking
- Admin moderation & audit trails
- 🔮 Pollution forecasting
- 🏭 Industrial compliance analytics
- 🏙️ Policy & urban planning insights
- 🧪 Research & startup APIs
- 💧 Smart purifier & IoT integration
From raw data → intelligence → prevention.
- Frontend: React.js, Tailwind CSS
- Visualization: Recharts, Leaflet
- Design Focus: UX for researchers
- Backend & AI: Conceptual (proposed)
This project is a proof-of-concept frontend prototype built to demonstrate how an AI-driven environmental intelligence system can reduce manual effort, minimize errors, and unlock the true value of research data.
⭐ Built for Smart India Hackathon (SIH)
- 📘 Detailed Project Documentation
👉 View Project Report
👉 Watch on YouTube