EnforceSmart is an AI-powered Smart City solution developed for the Flipkart Grid Hackathon to address parking-induced traffic congestion in Bengaluru.
Bengaluru is one of India's fastest-growing metropolitan cities and experiences severe traffic congestion due to rapid urbanization, increasing vehicle ownership, and illegal roadside parking. These issues lead to longer travel times, higher fuel consumption, delayed emergency services, and increased pollution.
EnforceSmart leverages Artificial Intelligence, Machine Learning, and Geospatial Analytics to analyze historical parking violation data from Bengaluru and provide predictive insights. Instead of reacting after congestion occurs, the system predicts high-risk parking hotspots, estimates congestion risk, and recommends optimal police patrol deployment to support proactive traffic management.
The platform enables traffic authorities to make data-driven decisions through an interactive dashboard featuring hotspot maps, heatmaps, violation analytics, and tactical patrol recommendations.
Bengaluru's roads face significant traffic congestion, with illegal and improper parking being one of the major contributing factors. Conventional traffic enforcement is largely reactive, where action is taken only after violations have already disrupted traffic flow.
The objective of EnforceSmart is to develop an AI-driven predictive policing system that:
- Predicts parking-induced congestion hotspots
- Identifies high-risk locations before congestion worsens
- Assists traffic police in efficient patrol deployment
- Provides actionable insights through real-time visualizations
- Supports smarter and more efficient urban traffic management
#SOLUTION
EnforceSmart is an AI-powered Smart City solution designed to analyze parking violation data, predict congestion hotspots, and assist law enforcement agencies in making proactive patrol deployment decisions. By combining Machine Learning, Geographic Information Systems (GIS), and interactive visualizations, the system transforms historical parking violation records into actionable insights for smarter urban traffic management.
- Predictive parking violation hotspot detection
- Interactive map visualization of congestion-prone areas
- Hourly, junction-wise, and offence-wise analytics
- Smart patrol deployment recommendations
- Risk score prediction for parking-induced congestion
- Heatmap visualization of parking violations
- Modern and responsive React dashboard
- React.js
- Tailwind CSS
- Leaflet
- Plotly
- Axios
- FastAPI
- Python
- Pandas
- NumPy
- XGBoost
- LightGBM
- DBSCAN
- Scikit-learn
enforce_smart/
│
├── backend/
│ ├── api/
│ ├── models/
│ ├── services/
│ ├── saved_models/
│ └── main.py
│
├── frontend/
│ ├── src/
│ ├── components/
│ ├── pages/
│ └── assets/
│
├── dataset/
├── notebooks/
└── README.md
- Collect and preprocess parking violation data.
- Detect congestion hotspots using DBSCAN clustering.
- Predict congestion risk using XGBoost.
- Classify dominant vehicle and violation types.
- Serve predictions through FastAPI APIs.
- Visualize insights using the React dashboard.
| Endpoint | Description |
|---|---|
/predict/hotspots |
Predict parking hotspots |
/heatmap/violations |
Generate parking violation heatmap |
/stats/hourly |
Hourly violation statistics |
/stats/offence |
Offence distribution |
/stats/junctions |
Junction-wise analytics |
git clone https://github.com/ezhil-34/enforce_smart.git
cd enforce_smartcd backend
pip install -r requirements.txt
uvicorn main:app --reloadcd frontend
npm install
npm run devThe project uses anonymized parking violation data containing information such as:
- Latitude & Longitude
- Junction Name
- Vehicle Type
- Violation Type
- Police Station
- Date & Time of Violation
- Device ID