New Approaches to Historical Maps of Atlanta: Combining Geo-AI Feature Extraction and Community Engaged Research

Alexander Cors, Emory University

Georgia universities and libraries hold historic and thematic maps of Atlanta that are invaluable sources of information on the city's urban development, transportation networks, and land-use planning. While projects like ATLmaps.org provide access to almost 3,000 maps from the last 170 years, features such as roads and buildings were not digitized due to the time requirements. Manually tracing intricate road networks from historical maps can take several weeks. Over the past year, our team at Emory University has developed an AI-enabled workflow to address this issue. Our GeoAI project pulls from advances made in road extraction for satellite imagery and adapts these techniques to historical cartographic data, saving time and human effort. By partnering with community partners and developing transferable AI methods, we create public datasets, maps, and narratives that reveal how past decisions continue to influence present-day disparities and offer an adaptable model that other cities can utilize.

No extended abstract or paper available

 Presented in Session 21. Mapping Atlanta's Historical Geographies of Race, Class, and Justice I