An investigation of the use of Google Street View for identifying gentrification across diverse United States morphological city types

Google Street View (GSV) is a tool for measuring characteristics of the built environment and appropriate for analyzing change in neighborhoods longitudinally. We investigate whether using GSV to identify gentrification works across the spectrum of urban morphologies in the U.S. We first use Hirsch and Schinasi's (2019, A measure of gentrification for use in longitudinal public health studies based in the United States. Drexel University Urban Health Collaborative.) gentrification criteria to identify tracts that underwent gentrification between 2012 and 2019 in Boston, Cincinnati, and Phoenix. We then used Hwang's (2015, Gentrification, race, and immigration in the changing American city [Unpublished doctoral dissertation] Harvard University.) GSV-based gentrification index, originally used in Chicago, to measure gentrification in the built environment and look for alignment with gentrification using Hirsch & Schinasi's criteria. Alignment was best in Boston with a population density and built form closest to Chicago's. The process did not work as well for Cincinnati where gentrification was concentrated in only a few blocks within gentrifying tracts nor in Phoenix with its limited amount of up-scale high-density apartments/condominiums and numerous vacant parcels. We conclude that cities with different urban morphologies need different methods to detect evidence of gentrification in the built environment.