Primary competition visual

EY Urban Heat Island Challenge

$3 500 USD
16 days left
Earth Observation
Geospatial Analysis
412 joined
84 active
Starti
Dec 12, 25
Closei
Jan 16, 26
Reveali
Jan 23, 26
About

🌡️ Ground Temperature Data

Collected by CAPA Strategies using sensors on cars and bicycles.

Two cities:

  • Santiago, Chile (21,662 points – Jan 20, 2024)
  • Rio de Janeiro, Brazil (28,488 points – Jan 27, 2023)

Each point includes latitude, longitude, and near-surface temperature.

Temperatures converted into a UHI Index and three classes:

  • Low (≤0.98)
  • Medium (0.98–1.02)
  • High (≥1.02)

🏙️ Building Footprints

  • 3D building footprint data (area + height).
  • Useful for modeling urban density, a major driver of local heat.

🛰️ Satellite Data (Sentinel-2)

  • 10 m optical imagery from ESA’s Sentinel-2.
  • Provided with cloud-filtered mosaics & example notebooks.
  • Enables creation of indices like NDVI (vegetation) and NDBI (built-up areas).

🎯 Your Task

Use these combined datasets - temperature labels, building footprints, and satellite-derived features—to predict Urban Heat Island class (Low / Medium / High) for new locations in Freetown, Sierra Leone.

Files
Description
Files
Is an example of what your submission file should look like. The order of the rows does not matter, but the names of the "ID" and column headers must be correct.
Get started with this Python notebook and make your first submission.
This is the dataset you will apply your model to.
Challenge and Data Overview.
Python package dependencies.
This Python notebook demonstrates cloud filtering, generation of a median mosaic, and outputs a GeoTIFF file with 12 spectral bands and 3 spectral indices.
This Python notebook outputs GeoTIFF files for 3 locations.
Training and building footprint data.