Cyclone Gezani Building Damage Assessment for Toamasina

Microsoft AI for Good Lab mapped affected buildings in Toamasina, Madagascar, following Cyclone Gezani using Airbus satellite imagery and AI models. The dataset provides geospatial damage assessment data. It was last updated in April 2026.

Use Cases

• Classify building damage status using AI-generated labels from satellite imagery features.
• Analyze spatial patterns of cyclone impact by mapping affected building locations.
• Estimate infrastructure recovery needs by aggregating counts of damaged buildings.
• Validate AI damage assessment models by comparing predicted labels with ground truth data.

Strengths

• Data is derived from Airbus satellite imagery, a high-resolution source.
• Assessment was performed by Microsoft AI for Good Lab's specialized models.
• Dataset was updated in April 2026, indicating recent maintenance.

Limitations

• The total number of assessed buildings (row count) is unknown.
• The accuracy metrics of the AI damage classification are not provided.
• Data is limited to the single event of Cyclone Gezani in Toamasina.

Provenance

Source

Microsoft AI for Good Lab.

Collection Method

AI model analysis of Airbus satellite imagery.

Time Range

Period surrounding Cyclone Gezani event (specific dates unknown).

Freshness

Last updated April 2026.

Geography

Toamasina, Madagascar.