Taubman College researcher Xiaofan Liang and colleagues have created an innovative new model to identify neglected and abandoned urban properties, which could ultimately help city officials protect and support growth within their communities.
Addressing strategies for managing vacant properties is important for maintaining healthy communities, but the process of identifying these properties can be difficult. Many cities identify neglected structures and land by visually identifying and recording these properties or by relying on residents’ calls to a civic hotline to report problems with properties in their neighborhoods. Other cities use a machine learning model, but reports of this strategy emphasize the technology itself rather than the combination of human input, technology, and data.
Partnering with the city of Savannah, Georgia, the research team designed a model that integrated human expertise into a machine-learning model that more reliably identified vacant and abandoned properties in Savannah. Their findings are published in the Journal of Planning Education and Research.
“By using a collaborative technique that engages experts while using large datasets and statistical analysis, we not only improve the local government’s ability to identify vacant and abandoned properties but also help the experts reflect on their approaches and where human discretions should remain,” said Liang, lead author and assistant professor of urban and regional planning. “The result is a more reliable method for managing assets in a municipal setting.”
Liang’s co-authors were Brian Brainerd and Tara Hicks from Savannah’s Housing and Neighborhood Services Department and Clio Andris from the Georgia Institute of Technology.
Savannah, a historic city with a population of nearly 150,000, has a growing number of deteriorating or neglected residential properties. In 2019, 1,319 properties had code violations that indicated severe physical deterioration and 1,404 properties had at least three years of tax delinquent history.
