GlobalBuildingAtlas: 3D Models of 2.8B Buildings in the World on GitHub (www.heise.de)

A team from the Technical University (TU) Munich has launched the GlobalBuildingAtlas, an extensive open dataset featuring high-resolution 3D models of 2.75 billion buildings worldwide, available on GitHub. Leveraging satellite imagery from 2019, this dataset includes 97% of buildings represented with Level of Detail 1 (LoD1) models, allowing for efficient integration into large-scale computer models. This comprehensive mapping is unprecedented, surpassing previous databases and filling gaps in global building data, particularly in regions like Africa and South America. The significance of the GlobalBuildingAtlas extends to urban planning, social research, and disaster preparedness. Its high accuracy (3m × 3m resolution) and detailed insights can help identify social and economic disparities by enabling analyses such as per capita building volume. Urban planners are already considering its implications for effective housing and public facility planning in underserved areas. With growing interest from institutions like the German Aerospace Center (DLR), this dataset is poised to be a valuable resource for addressing urban challenges and fostering inclusive development.