Transforming data-dark zones into actionable intelligence using AI

🏆 4th place out of 100+ teams at InnovateX 4.0 (International Tech Fest @ Presidency)

Urban slums experience disproportionately high crime rates, yet a large portion of these incidents go unreported due to:

• Fear of retaliation
• Lack of anonymity
• Limited access to reporting systems
• Low digital literacy

This leads to data invisibility, where entire communities are excluded from official datasets.

❌ No data → ❌ No visibility → ❌ No intervention

Urban slums often function as data-dark zones, where crime exists but is not reflected in structured systems.

This creates critical challenges:

• 📉 Authorities rely on incomplete or biased data
• 🚫 Preventive measures are rarely implemented
• ⚠️ Crime patterns remain invisible

The real problem is not just crime — it is the absence of reliable data.

SlumSafe AI is a predictive and participatory crime intelligence system that:

• 🧠 Predicts crime risk using AI
• 🗺️ Visualizes hotspots via heatmaps
• 📢 Enables anonymous reporting
• 🚨 Provides emergency action support

🔥 We don’t just analyze crime — we create visibility where none exists

• Machine learning model (Random Forest)
• Uses time and location patterns
• Estimates crime risk even with limited data

• Interactive heatmap using Folium
• Converts predictions into intuitive insights
• Seamless "Quick Jump" navigation across global hotspots (Mumbai, Bangalore, Goa)

• Instant 1-Tap Anonymous reporting system (No typing required)
• Encourages community contribution
• Reduces underreporting

Limited Data → Prediction → User Reports → More Data → Better Predictions

• Real-time alerts
• Emergency contact feature

This UI is a conceptual design created by our team to visualize the system before implementation. It reflects our planned workflow and user experience for the final product.

• Inputs: Latitude, Longitude, Time
• Output: Risk Level (Low / Medium / High)

• Color-coded risk zones

  • 🔴 High
  • 🟡 Medium
  • 🟢 Low

• Instant 1-Tap interface for reporting incidents dynamically
• Automatically captures and maps location/timestamp directly to heatmap
• Stores data locally (CSV)

• One-click emergency call
• Uses device dialer (tel: link)

• Highlights high-risk areas
• Time-based warnings

User Input (Location + Time) 
        ↓
Data Processing
        ↓
ML Model Prediction
        ↓
Risk Classification
        ↓
Heatmap Visualization + Alerts
        ↓
1-Tap Anonymous Reporting
        ↓
Continuous Improvement

SlumSafe-AI/
│
├── app.py
├── requirements.txt
├── README.md
├── LICENSE
├── UI-preview.jpeg
│
├── scripts/
│   ├── crime_model.py
│   ├── fetch_chicago.py
│   └── gen_india_data.py
│
├── model/
│   └── model.pkl
│
└── data/
    ├── crime_data.csv
    ├── ngos.csv
    ├── emergency_contacts.csv
    └── reports.csv

git clone https://github.com/gee-46/SlumSafe-AI.git
cd SlumSafe-AI
pip install -r requirements.txt
streamlit run app.py

• 🚔 Enables proactive policing
• 🏙️ Supports urban planning
• 🤝 Helps NGOs target interventions
• 👥 Empowers underserved communities

• Uses simulated data for prototyping
• Model accuracy improves over time

• Real-time data integration
• SMS-based reporting
• NGO / police API integration
• Advanced analytics

Event        : InnovateX 4.0 (International Tech Fest @ Presidency)
Track        : Build With AI (24-Hour Hackathon)
Team         : PulseX
Result       : 🏅 Secured 4th Place
Competition  : 100+ Teams

PulseX

• Aniroshgouda Ramanagoudar (Team Leader)
• Gautam N Chipkar
• Shridharsingh Rajaput
• Basavaraj Basagaudar

• Modular development using Git branches
• Parallel development with structured integration
• Focus on rapid prototyping and clean execution

“Built with collaboration, speed, and impact in mind.”

If you found this project interesting:

⭐ Star the repo 🍴 Fork it 💡 Build on it

“We are not just predicting crime — we are making invisible communities visible in data systems.”