mwBTFreddy is a dataset with images specific to Malawi arranged in the format of xBD.

xBD is a global building damage assessment dataset containing pre- and post-disaster satellite images with labeled building footprints and damage classifications.

Test Dataset:

• Malawi-Specific Dataset: A dataset curated by the Kuyesera AI Lab, with hand-labeled pre- and post-disaster images from Cyclone Freddy. This includes:
• Pre-Disaster Imagery: Captured approximately 6 months before the cyclone.
• Post-Disaster Imagery: Captured after Cyclone Freddy’s landfall.

ASDI Datasets (Required for Prize Eligibility):

• You must use at least one dataset from the Amazon Sustainability Data Initiative (ASDI) platform as part of your model. Recommended starting points include:
• Climate and Weather Data: To incorporate meteorological patterns.
• Land Cover and Vegetation Data: To account for environmental factors like deforestation.

Potential starting datasets

• Soil info dataset - https://aws.amazon.com/marketplace/pp/prodview-z6d6q7p7a65xe?sr=0-13&ref_=beagle&applicationId=AWSMPContessa#usage
• Elevation models -https://aws.amazon.com/marketplace/pp/prodview-yz6tvt6alc5ny?sr=0-19&ref_=beagle&applicationId=AWSMPContessa#resources / https://aws.amazon.com/marketplace/pp/prodview-yz6tvt6alc5ny?sr=0-19&ref_=beagle&applicationId=AWSMPContessa#resources
• S2 cloud optimised images - https://aws.amazon.com/marketplace/pp/prodview-ykj5gyumkzlme?sr=0-29&ref_=beagle&applicationId=AWSMPContessa
• SAR data - https://aws.amazon.com/marketplace/pp/prodview-ybjvekte4hg7o?sr=0-37&ref_=beagle&applicationId=AWSMPContessa

You can download the data for the competition from the following links:

• https://kuyesera-ai-challenge.s3.amazonaws.com/test_image_coords.csv
• https://kuyesera-ai-challenge.s3.amazonaws.com/test_images.zip
• https://xview2.org/dataset (registration required)