Explainable AI · SHAP-based decomposition

Driver Analysis

For every hotspot, HeatPilot decomposes the heat score into physical drivers — built-up density, vegetation, wind, humidity, morphology — using a physics-informed model.

Shivajinagar

Ward 12 · LST 44.8°C

Extreme
Heat Score
94/100
Built-up Density42%
Vegetation Loss31%
Road Network14%
Humidity8%
Wind5%
Top driver is Built-up Density. Targeted intervention can yield −2.8°C local cooling.
Hadapsar

Ward 07 · LST 43.6°C

Very High
Heat Score
85/100
Built-up Density38%
Population Density24%
Vegetation Loss22%
Road Network11%
Wind5%
Top driver is Built-up Density. Targeted intervention can yield −2.5°C local cooling.
Kothrud

Ward 21 · LST 42.1°C

Very High
Heat Score
75/100
Built-up Density34%
Road Network22%
Vegetation Loss20%
Humidity14%
Wind10%
Top driver is Built-up Density. Targeted intervention can yield −2.3°C local cooling.
Yerwada

Ward 15 · LST 43.2°C

Very High
Heat Score
82/100
Built-up Density36%
Vegetation Loss26%
Population Density18%
Road Network12%
Wind8%
Top driver is Built-up Density. Targeted intervention can yield −2.4°C local cooling.
Kasba Peth

Ward 03 · LST 42.7°C

Extreme
Heat Score
79/100
Built-up Density46%
Vegetation Loss28%
Urban Morphology14%
Humidity8%
Wind4%
Top driver is Built-up Density. Targeted intervention can yield −3.1°C local cooling.
Kondhwa

Ward 18 · LST 41.4°C

High
Heat Score
70/100
Built-up Density30%
Road Network24%
Population Density20%
Vegetation Loss16%
Wind10%
Top driver is Built-up Density. Targeted intervention can yield −2.0°C local cooling.
HeatPilot · Developed by Team PRAGYAN · Bharatiya Antariksh Hackathon 2026 · ISRO Problem Statement 1
AI-Powered Decision Support for Urban Heat Mitigation