Microsoft AI for Good Lab

Member since 8 January 2025
Data Datasets [3] | Archived Datasets[0] [?]
Refine your search: Clear all
Featured:
Locations:
Formats:
Organisations:
Tags:
Licenses:
  • LA Fires - Palisades fire January 2025: Building Damage Assessment
    Microsoft AI for Good Lab
    30+ Downloads
    Time Period of the Dataset [?]: January 08, 2025-January 10, 2025 ... More
    Modified [?]: 14 January 2025
    Dataset Added on HDX [?]: 9 January 2025
    Less
    This dataset updates: As needed
    We ran our damage assessment AI models on images provided by Planet and Maxar and have mapped out the affected buildings. The images cover Palisades fires. Analysis was done on: 1) Planet's 20250108_222134_ssc10 SkySat scene captured 01/08 2) Planet's 20250109_221527_ssc10 SkySat scene captured 01/09 3) Maxar imagery 1050010040277500 captured 1/10 Latest results show the following. Total of 18,538 structures: 11,735 not damaged 6,803 damaged
  • LA Fires - Eaton fire January 2025: Building Damage Assessment
    Microsoft AI for Good Lab
    Time Period of the Dataset [?]: January 10, 2025-January 11, 2025 ... More
    Modified [?]: 14 January 2025
    Dataset Added on HDX [?]: 11 January 2025
    Less
    This dataset updates: Never
    We ran our damage assessment AI models on images provided by Maxar and Planet and have mapped out the affected buildings. The images cover Eaton fire. Analysis was done on: 1) Maxar imagery captured 01/10 2) Planet imagery captured 1/11 Latest results show the following. 156,102 buildings in study area 8,682 estimated damaged
  • Cyclone Chido: Building Damage Assessment in Passamanty, Mayotte
    Microsoft - AI4G Lab
    Time Period of the Dataset [?]: December 09, 2024-December 09, 2024 ... More
    Modified [?]: 8 January 2025
    Dataset Added on HDX [?]: 8 January 2025
    Less
    This dataset updates: Never
    We ran our damage assessment AI models on images provided by Planet and have mapped out the affected buildings. The images cover Passamainty in Mayotte. We assessed 3,875 buildings in the area, with damage levels as follows: 1,426 buildings are between 0% and 20% damaged 136 buildings are between 20% and 40% damaged 92 buildings are between 40% and 60% damaged 62 buildings are between 60% and 80% damaged 2,159 buildings are between 80% and 100% damaged