DrivenData is hosting Open Cities AI Challenge: Mapping Buildings for Disaster Resilience now through March 16, 2020.

Urban populations are rapidly growing across the globe. In Africa, populations are expected to double in the next 25 years. It’s important to understand the built environment of a city (buildings, roads, and critical infrastructure) in order to foster resilience to natural hazards and climate change. To do this well, cities should be mapped frequently, accurately, and with a high level of detail.

There is an increasing amount of labeled imagery data in OSM that could be used to address this issue. The goal of the challenge is to utilize this data to develop better open-source building segmentation models and apply machine learning systems in the most responsible way in order to manage disaster risk in African cities. Better performing and responsibly used AI systems can provide more accurate, faster, and lower-cost approaches to assessing risk and protecting lives and property.

Azavea partnered with DrivenData and GFDRR Labs to create the data for this competition. $15,000 will be awarded in prizes across two different tracks. The competition is live now through March 16, 2020. Register for an account and start building your models!