urbanCool

What Inspired Us

The inspiration for UrbanCool stems from the growing climate crisis and its disproportionate impact on urban populations. As global temperatures rise, the "Urban Heat Island" effect is no longer just a scientific phenomenon—it is a public health emergency. We were moved by stories of cities where a lack of tree canopy in low-income neighborhoods led to significantly higher temperatures and energy costs compared to wealthier, greener areas. We wanted to build a tool that democratizes access to high-level satellite data, turning complex climate science into a simple, actionable map for city planners and community advocates.

What We Learned

Developing this project was a deep dive into the intersection of Remote Sensing and Web Development. We learned that:

• Data Fusion is Key: Surface temperature from satellites like Landsat is powerful, but combining it with high-resolution reflectance data from Sentinel-2 is what allows us to see heat at a "street-level" resolution (20m ).

• Accessibility Matters: Complex geospatial data is useless if it's trapped in a research paper. The real challenge isn't just mapping the heat; it's designing a UI that tells a planner exactly where to plant a tree to save the most lives.

• The Power of Open Source: The wealth of data provided by NASA and the ESA, combined with open-source tools like PostGIS and MapLibre, makes it possible for a small team to build tools that were once only available to national space agencies.

How We Built It

UrbanCool was built using a modular, data-first approach:

1. Data Pipeline: We leveraged Google Earth Engine to handle the heavy lifting of processing petabytes of satellite imagery. We created scripts to filter for cloud-free days and calculate Land Surface Temperature (LST) and NDVI (Vegetation Index).

2. Geospatial Backend: We used FastAPI and PostGIS to serve this data. Instead of sending massive raw files to the user, we converted the data into Vector Tiles, allowing for smooth, real-time interaction on the map.

3. Interactive Frontend: The UI was crafted with React and Mapbox GL JS. We focused on a "Thermal-First" design, ensuring that the heatmap overlay was intuitive and that the "Actionable Insights" panel provided clear, data-driven recommendations.

4. Cloud Infrastructure: The entire system is hosted on AWS, utilizing S3 for tile storage and Lambda for on-the-fly spatial queries, ensuring the app remains fast and scalable as more cities are added.

Built With

• apis
• figma
• manus
• playground
• stitch
• windsurf