Built for the National Road Safety Hackathon 2026 Organized by Centre of Excellence for Road Safety (CoERS), IIT Madras
🌐 Live Demo: https://roadwatchai.streamlit.app/
📂 GitHub Repository: https://github.com/IamDip-SK10/RoadWatch_AI
RoadWatch AI is an intelligent road safety analytics and infrastructure transparency platform designed to empower citizens, engineers, and policymakers with data-driven insights.
The platform combines:
- 📊 Advanced road accident analytics
- 🗺️ Geospatial hotspot identification
- 🤖 AI-powered road safety assistant
- 📋 Automated grievance generation
- 🏗️ Contractor transparency tracking
- 💰 Infrastructure budget visibility
- 🏛️ Executive Engineer routing system
RoadWatch AI directly aligns with the IIT Madras RoadWatch challenge theme by promoting transparency, accountability, and citizen engagement in road infrastructure governance.
Interactive dashboard powered by Plotly visualizations:
- Monthly accident trends
- Hourly accident distribution
- Day-wise accident analysis
- Peak vs non-peak casualty analysis
- Severity distribution
- Environmental factor analysis
- Risk profiling
- State-wise casualty monitoring
RoadWatch AI identifies accident-prone regions using:
- Geographic hotspot mapping
- Risk density visualization
- State-level accident clustering
- Location-based risk scoring
This helps authorities prioritize road safety interventions and infrastructure investments.
Integrated with Google Gemini 1.5 Flash.
Capabilities include:
- Accident trend summaries
- Risk hotspot analysis
- Grievance guidance
- Road safety insights
- Infrastructure transparency queries
The assistant uses a hybrid architecture:
- Dataset-driven local responses
- Gemini-powered intelligent responses
- Rule-based fallback engine
This ensures uninterrupted service even when AI APIs are unavailable.
Citizens can:
- Select one or multiple cities
- Generate official grievance tickets
- View responsible authorities
- Access infrastructure records
- Submit road safety complaints
Each ticket includes:
- Unique reference number
- Executive Engineer details
- Department information
- Budget allocation records
- Maintenance history
- Contractor information
A core innovation of RoadWatch AI.
For every selected city, users can view:
- Assigned contractor
- Infrastructure budget
- Last road relaying date
- Responsible department
- Executive Engineer information
This increases accountability and transparency in public infrastructure projects.
The platform enables monitoring of:
- Road maintenance budgets
- Infrastructure allocations
- City-wise spending visibility
- Budget accountability tracking
This bridges the transparency gap between public funds and infrastructure outcomes.
Dataset Upload
│
▼
Data Validation Layer
│
▼
Analytics Engine
│
┌────┴────┐
▼ ▼
Dashboard AI Assistant
│ │
▼ ▼
Risk Maps Gemini Integration
│ │
└────┬─────┘
▼
Grievance Routing Engine
│
▼
Infrastructure Transparency Layer
| Layer | Technology |
|---|---|
| Frontend | Streamlit |
| Analytics | Pandas |
| Numerical Processing | NumPy |
| Visualization | Plotly |
| AI Assistant | Google Gemini 1.5 Flash |
| Mapping | Plotly Mapbox |
| Data Storage | CSV / ZIP Datasets |
| Deployment | Streamlit Community Cloud |
| Package | Purpose |
|---|---|
| streamlit | Web application framework |
| pandas | Data manipulation |
| numpy | Numerical processing |
| plotly | Interactive visualizations |
| google-generativeai | Gemini AI integration |
| zipfile | ZIP dataset support |
| io | Memory-based file handling |
| datetime | Date operations |
| random | Deterministic infrastructure generation |
The application supports datasets containing:
- Accident records
- Geographic coordinates
- Road types
- Severity levels
- Weather conditions
- Visibility metrics
- Traffic density
- Casualty information
- Risk scores
Supported formats:
- CSV
- ZIP containing CSV
RoadWatch AI uses a three-layer intelligence model:
Answers generated directly from filtered datasets.
Complex questions routed to Gemini 1.5 Flash.
If Gemini is unavailable:
- No crash
- No interruption
- Automatic fallback responses
The project directly addresses the official RoadWatch theme:
✅ Road Quality Monitoring
✅ Public Infrastructure Transparency
✅ Citizen Grievance Management
✅ Contractor Accountability
✅ Budget Visibility
✅ Road Safety Analytics
✅ Data-Driven Governance
✅ AI-Assisted Decision Support
- Real-time government API integration
- Mobile application
- Image-based road damage detection
- Predictive accident forecasting
- Smart contractor performance scoring
- Live infrastructure monitoring
- Citizen reputation system
- GIS-based route safety recommendation
Subhadip Kumar
Team: Team Subhadip
Role:
- Full Stack Development
- Data Analytics
- AI Integration
- Visualization Design
- Documentation
- Deployment
Built independently as a solo submission for the National Road Safety Hackathon 2026.
MIT License
This project is provided for educational, research, and hackathon purposes.