Most low-and middle-income countries are experiencing the proliferation of deprived areas within urban spaces. Such areas are prone to climate-related hazards and poverty. This study map deprived areas using open geospatial data and machine learning. The study conceptualized deprived areas, develop a consistent feature-creation method for integrating varying spatial and temporal resolutions, train machine learning models and evaluated on unseen geographic regions. OSM data were used for preparing training data and spatial indicators. Models achieve an F1 score 65% - 0.83% for all the three cities. This demonstrates the potential of open geospatial and machine learning to predict deprived areas in unseen geographic areas.
Speaker
Next up in State of the Map US
Previous talkMVT Tiles and Autonomous Driving, and other non-UI usage
Map Vector Tiles can be used for more than just drawing maps. We will discuss community and industry-suggested MVT usage, and how we can grow the standard beyond the initial visualization use...
