Examining state-led gentrification using street view imagery: Evidence from Shanghai, China

This study introduces a transferable framework for detecting gentrification by leveraging detailed street-view imagery to capture visible changes in the built environment at the street level. The research utilizes deep learning techniques to analyze paired images from Shanghai spanning 2013 to 2022, allowing for the identification of neighborhood-level improvements. These visual findings are then systematically aligned with census data and case studies to strengthen the analysis. The approach proves especially effective within the context of state-orchestrated, property-led redevelopment, which is characterized by densification, public-space enhancement, and functional reconfiguration—hallmarks of entrepreneurial urbanism. The results reveal that such physical and functional upgrades are concentrated in central districts, waterfront corridors, and selected new towns, demonstrating that gentrification in Shanghai has gone beyond the traditional inner city, expanding to amenity-rich, brandable environments. The framework translates landscape transformations into consistent, citywide evidence, complementing conventional demographic and price-based measures. It offers spatiotemporally fine-grained monitoring and holds potential for application in other rapidly changing cities experiencing state-led gentrification.