21,994 images for binary damage classification (damage vs. no damage) across four natural disaster events. The dataset was created by author 'abalhomaid' and is hosted on Hugging Face. It covers the 2016 Ecuador Earthquake, 2015 Nepal Earthquake, 2016 Hurricane Matthew, and 2014 Typhoon Ruby.
Use Cases
- Training binary damage classifiers based on satellite or aerial imagery.
- Benchmarking domain adaptation methods across different disaster types mentioned in the description.
- Developing rapid assessment tools for post-disaster response based on the damage classification task.
Strengths
- Contains 21,994 labeled images across four distinct disaster domains.
- Provides a clear binary classification task (damage vs. no damage) for model training.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count per domain is provided, but total row count for the structured dataset is unknown.
- Last updated 2026-04-12 05:03:07; freshness should be verified.
Provenance
- Source
- Hugging Face (author: abalhomaid).
- Collection Method
- Likely compiled from satellite or aerial imagery sources for specific disaster events.
- Time Range
- Covers events from 2014 to 2016.
- Freshness
- Last updated 2026-04-12 05:03:07.
- Geography
- Covers Ecuador, Nepal, and regions affected by Hurricane Matthew and Typhoon Ruby.