Network Entropy as a Measure of Socioeconomic Segregation in Residential and Employment Landscapes (2304.10300v1)

Abstract: Cities create potential for individuals from different backgrounds to interact with one another. It is often the case, however, that urban infrastructure obfuscates this potential, creating dense pockets of affluence and poverty throughout a region. The spatial distribution of job opportunities, and how it intersects with the residential landscape, is one of many such obstacles. In this paper, we apply global and local measures of entropy to the commuting networks of 25 US cities to capture structural diversity in residential and work patterns. We identify significant relationships between the heterogeneity of commuting origins and destinations with levels of employment and residential segregation, respectively. Finally, by comparing the local entropy values of low and high-income networks, we highlight how disparities in entropy are indicative of both employment segregation and residential inhomogeneities. Ultimately, this work motivates the application of network entropy to understand segregation not just from a residential perspective, but an experiential one as well.