🏆 1st Place Winner – Vashisht 2024 Data Analytics Hackathon
NexusAeon Data Analytics Hackathon is a data-driven road safety analysis hackathon that leverages over 1.8 million real-world driving events from ADAS (Advanced Driver Assistance Systems) and DMS (Driver Monitoring Systems) sensors to identify accident blackspots and risky driver behaviors. Our solution combines geospatial analytics, clustering, time-based visualizations, and domain-specific road safety references to suggest actionable safety interventions.
Presented with an interactive Power BI dashboard and a concise data story, this project secured the 1st position at the Vashisht 2024 Hackathon, hosted by IIITDM Kancheepuram.
- Mithilesh Gopalakrishnan S
- Shobhan Karthish M
Given telemetry and alert data from 130 vehicles collected over 30 days (spanning multiple Indian states), identify:
- High-frequency accident zones
- Risky timeframes for alerts
- Driving behavior patterns (e.g., drowsiness, unsafe distances, mobile usage)
- Actionable interventions for improving road safety
- Vehicles: 130 fleet vehicles
- Duration: 30 days
- Events: 1.8+ million alerts and sensor readings
- Regions: Karnataka, Telangana, Andhra Pradesh, Maharashtra, Odisha, Tamil Nadu, Kerala
- Alert Types:
- FCW: Forward Collision Warning
- PCW: Pedestrian Collision Warning
- LDW: Lane Departure Warning
- HMW: Headway Monitoring Warning
- DMS Alerts: Drowsiness, Seatbelt, Mobile Usage, Yawning
- Removed noisy/missing points
- Focused on events with high confidence values
- Used heatmaps and DBSCAN to identify blackspot clusters
- Mapped alerts over road networks and accident-prone areas
- Created hourly and daily time-series graphs
- Donut charts and line plots to visualize alert spikes
- Grouped alerts by type: CAS (collision) vs. DMS (driver behavior)
- Separated analytics for each group
| 📍 Location | 🚨 Incidents | 🔎 Observation |
|---|---|---|
| Wardha Road, Nagpur | 30,000+ | Poor visibility, high-speed segment |
| HYD–VIZAG Highway (NH-65) | 12,000+ | Multiple DMS alerts – fatigue & mobile use |
| Airport Rd, Karnataka (NH-43) | 10,000+ | Unsafe lane changes, low signage |
Additional DMS alert hotspots were seen along urban congested regions like Dilsukhnagar and Mahatma Gandhi Bus Station Area (HYD).
The project included a complete Power BI dashboard with:
- Dynamic filters (alert type, time range, location)
- Alert frequency graphs
- Geospatial heatmaps
- Blackspot ranking dashboard
- Alert trend by hour/day/week
📂 See visualizations/ for exported views or the .pbix file.
| Problem | Recommendation |
|---|---|
| Frequent CAS alerts at highways | Speed calming zones, better signboards |
| High DMS alerts (drowsiness, mobile) | Awareness campaigns, stricter traffic enforcement |
| Poor lane management | Road marking improvement and lane sensors |
| Bad road surfaces | Infrastructure maintenance and pothole patching |
- Road Safety Audit: Wardha Road – Nagpur City
- Road Accidents in Maharashtra 2020 – NH65 Case Study
- Times of India: NH-65 – Bloodiest Stretch in Telangana
NexusAeon-Vashisht-Hackathon/
├── report/
│ └── Vashisht_2024_NexusAeon.pdf
├── visualizations/
│ └── dashboard_screenshots.png
│ └── powerbi_dashboard.pbix
├── README.md