European Urban Chronic Heatwave Spatial Exposure & Resilience Predictor
Escalating heatwaves in European cities present chronic infrastructure and health risks. Build a spatial ML engine using GeoPandas, Rasterio, and Prophet time-series models to predict urban heat island intensity and quantify resilience infrastructure investment effectiveness.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a geospatial ML pipeline combining GeoPandas and Prophet to forecast urban heat risks and evaluate resilience investments.
How work is evaluated
Evaluates spatial urban heat island prediction accuracy and resilience investment benefit estimation.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
positive_reduction
Verifies temperature reduction is positive for canopy increases
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
spatial_temp_mae
Mean Absolute Error of land surface temperature spatial prediction • target: 0.4 • range: 0-2
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Process thermal satellite raster imagery (e.g. Landsat surface temperature) with Rasterio
Build spatial heat island exposure maps using GeoPandas vector layers
Train Prophet models to forecast summer urban extreme temperature frequencies
Quantify expected municipal loss reductions from urban canopy greening investments
Reference links and supporting material
European city land surface temperature raster data paired with urban district vector shapefiles.
Challenge setup
Evaluates spatial urban heat island prediction accuracy and resilience investment benefit estimation.
Challenge input
JSON with district_id, baseline_canopy_pct, proposed_canopy_pct, and geographic_centroid.
GeoPandas
Vector manipulation and spatial analysis
Prophet
Seasonal trend forecasting for climate temperature time-series
Evaluated output
JSON with predicted_temp_reduction_c, estimated_health_cost_savings_eur, and resilience_score.
- Verifies temperature reduction is positive for canopy increases
- Mean Absolute Error of land surface temperature spatial prediction • target: 0.4 • range: 0-2
- Benchmark: UrbanHeatBench
- Spatial Temp Mae target: 0.4
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Prophet · Required
- xAI · Optional
- Baseten · Optional
Evaluation contract
- positive_reduction · Weight 1
- spatial_temp_mae · Weight 1
Metrics
- spatial_temp_mae · Target 0.4
Benchmarks
- UrbanHeatBench
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with district_id, baseline_canopy_pct, proposed_canopy_pct, and geographic_centroid.
Output: JSON with predicted_temp_reduction_c, estimated_health_cost_savings_eur, and resilience_score.
Execution harness: Python
Evaluation tools
- GeoPandas · Geospatial Data Processing
- Prophet · Time-Series Modeling
Recipe freeze marker
No recipe freeze time is recorded.