Climate-aware pedestrian routing for Brownsville, Brooklyn.

pip install -r requirements.txt

The app uses a local LLM via Ollama. Install Ollama and pull a model:

# Install Ollama: https://ollama.com/download
ollama pull qwen2.5:3b

streamlit run app.py

The app will open at http://localhost:8501.

deployment_package/
├── app.py                          # Main Streamlit application
├── requirements.txt                # Python dependencies
├── README.md                       # This file
├── core/
│   ├── __init__.py                # Module exports
│   ├── config.py                  # UI configuration
│   ├── engine.py                  # Routing engine (climate-aware weights)
│   └── tools.py                   # Routing tools and geocoding
└── data/
    ├── tool_embeddings.npz        # Pre-computed embeddings for tool selection
    └── brownsville/
        ├── graph_cache.pkl         # Cached graph (fast loading)
        ├── walking_network_final.graphml  # Full graph with climate attributes
        ├── all_resources.csv       # POI database (cooling centers, hospitals, etc.)
        ├── places.csv              # Geocoding database for place names
        └── pois_metadata.json      # POI type definitions

• Climate-Aware Routing: Routes avoid flood zones and high heat areas
• Accessibility Support: Grade factor penalizes steep hills for mobility-impaired users
• Natural Language Interface: Ask questions like "Find the nearest pharmacy" or "Route to Brookdale Hospital avoiding flooding"
• Multiple Route Alternatives: Shows shortest, flattest, and safest routes

weight = length * flood_mult * heat_mult * shade_mult * aqi_mult * grade_mult

The LLM adjusts these parameters based on user context:

• Flooding mentioned: flood_penalty_deep=10.0
• Heat/shade needed: heat_factor=0.5, shade_factor=0.5
• Elderly/wheelchair/mobility: grade_factor=0.5
• Respiratory concerns: aqi_factor=0.5

• Python 3.10+
• Ollama running locally with a compatible model
• ~15MB disk space for data files

"Could not connect to Ollama": Make sure Ollama is running (ollama serve)

Model not found: Pull the model first (ollama pull qwen2.5:3b)

Import errors: Ensure all dependencies are installed (pip install -r requirements.txt)