Exploring Urban Shrinkage via Computational Approaches: A Case Study of the City of Detroit



Journal Title

Journal ISSN

Volume Title



While the world’s total urban population continues to grow, not all cities are witnessing such growth, some are actually shrinking. This shrinkage causes several problems to emerge including population loss, economic depression, vacant properties and the contraction of housing markets. Such problems challenge efforts to make cities sustainable. While there is a growing body of work on study shrinking cities, few explore such a phenomenon through computational approaches. To further explore on this phenomenon, this work mainly focuses on three main questions: RQ1:How does urban shrinkage emerge at the macro-level through the simulation of housing trades at the individual level? RQ2: How can patterns of shrinkage be measured through a simulation? RQ3: To what extent can urban shrinkage be revealed by the analysis of newspaper articles? This work uses Detroit Tri-County Area as a study area to explore these three research questions. Two agent-based models (ABM) are built to simulate the housing trades within the Detroit Tri-County Area and the two models' results capture the city of Detroit's shrinkage from the aspect of decreasing numbers of households at the macro-level, which can be considered as the population loss. In addition, a pattern with a collapsing downtown housing market can be measured by visualizing the simulation results. By utilizing natural language processing (i.e., topic modeling) on a large number of newspapers related to Detroit, insights related to Detroit's shrinkage can be linked to the side effects of the economy recession on Detroit's automobile industry, local employment status, and housing market. The two agent-based models built in this work significantly add to the nascent field of inquiry by specifically capturing how the buying and selling of houses can lead to urban shrinkage from the bottom-up, which contributes to the computational social science (CSS) field of social simulation modeling. The topic modeling extends the field of text-contents analysis by identifying the insights related to urban shrinkage phenomenon from a large number of newspapers, which adds to the field of social information extraction under the CSS domain.