Targeted Urban Tree Planting to Reduce Heat Inequality

An interactive atlas that shows where tree planting can most reduce urban heat exposure using heat, tree, and demographic data.

Comment

Inspiration

Urban heat is not evenly distributed—some neighborhoods experience significantly higher temperatures than others, often overlapping with lower-income communities. We wanted to build a tool that helps cities answer a simple but important question: where should we plant trees to reduce heat exposure for the most people?

What it does

We built an interactive “Opportunity Atlas” for urban heat. The tool identifies neighborhoods where tree planting would have the greatest impact on reducing heat exposure by combining:

  • Urban heat data
  • Tree coverage data
  • Population and poverty data

The result is a priority map that highlights high-impact areas for intervention, while clearly distinguishing non-urban regions and areas without data.

How we built it

We used Python and GeoPandas in Google Colab to clean and merge multiple spatial datasets, including:

  • Satellite-derived tree data (2022)
  • Urban Heat Index data
  • U.S. Census population and poverty data

We performed spatial joins, created density metrics, and built a priority score that balances heat, tree coverage, and population exposure. The data was exported into a unified GeoJSON file and visualized using Leaflet and D3 to create an interactive atlas-style dashboard.

Challenges we ran into

One major challenge was data alignment across different spatial formats and resolutions. Heat and tree data did not always align perfectly, leading to missing values in some regions.

We also encountered edge cases like large non-urban areas (e.g., national forests), where standard metrics produced misleading results. We addressed this by explicitly identifying and separating non-urban regions from the analysis.

Accomplishments that we're proud of

  • Built a fully interactive atlas-style dashboard from scratch
  • Successfully integrated multiple geospatial datasets into a single pipeline
  • Designed a clear and interpretable priority scoring system
  • Identified and handled real-world data issues like missing coverage and non-urban regions

What we learned

We learned how to work with complex geospatial data pipelines and how important it is to critically evaluate data quality. We also saw firsthand how small modeling decisions—like handling missing data or excluding non-urban areas—can significantly impact results.

What's next

Next, we would:

  • Improve heat data coverage using additional datasets
  • Incorporate more environmental variables (e.g., land surface temperature, vegetation indices)
  • Expand the tool to other cities
  • Turn this into a real decision-support tool for urban planners and policymakers

Built With

Share this project:

Updates