HeatWatch

Open-source urban heat island intelligence for climate adaptation Map heat risk, vulnerable communities, and cooling potential using satellite data

MIT License · Python 3.10+ · NASA Earth data

---

Why this exists

Cities are getting hotter. Urban Heat Islands raise temperatures by 2–10°C, increasing mortality and inequality.

Most cities still lack:

• High-resolution heat maps
• Neighborhood-level vulnerability data
• Tools to evaluate cooling interventions

HeatWatch makes this accessible using open satellite + census data.

---

What it does

• Generates Land Surface Temperature (LST) heat maps from Landsat data
• Computes vegetation cover (NDVI) to estimate cooling capacity
• Scores neighborhood-level heat vulnerability
• Tracks heat change over time
• Exports GIS-ready outputs (GeoJSON, CSV, PNG)

---

Quick start

git clone https://github.com/your-org/heatwatch.git
cd heatwatch

python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

pip install -e ".[dev]"

heatwatch analyze --sample-city detroit

---

Example output

Hottest:  West Pullman   (42.1°C)
Coolest:  Lincoln Park   (31.8°C)
Risk:     Englewood      (high heat + low tree cover)

Saved → ./results/detroit/

---

Features

• Satellite-based heat mapping (Landsat)
• Vulnerability scoring (population + heat exposure)
• Vegetation analysis (NDVI)
• Multi-year comparisons
• GIS-ready exports

---

Tech stack

• Python 3.10+
• rasterio, geopandas
• NASA Earthdata / USGS APIs
• FastAPI
• folium
• pytest

---

Project structure

heatwatch/
├── src/
├── data/
├── tests/
├── docs/
└── ROADMAP.md

---

Docker (1 command setup)

Run the entire project without installing Python or dependencies:

docker build -t heatwatch .
docker run --rm -it heatwatch heatwatch analyze --sample-city detroit

---

Roadmap

• MVP: LST + NDVI + CLI
• Phase 2: Web dashboard + more cities
• Phase 3: Global dataset + ML-based prediction

---

Contributing

Start here:

1. Pick a “good first issue”
2. Fork repo
3. Create branch
4. Submit PR

We especially need help with:

• New city data support
• Visualization improvements
• Performance optimization
• Tests and documentation

---

License

MIT — free to use, modify, and distribute.

---

Contact

Issues: https://github.com/your-org/heatwatch/issues Discussions: https://github.com/your-org/heatwatch/discussions

---

Docker file (required to make the 1-command setup work)

Add this as Dockerfile:

FROM python:3.10-slim

WORKDIR /app

COPY . .

RUN pip install --no-cache-dir -e ".[dev]"

ENTRYPOINT ["heatwatch"]