Uneven trajectories of urban change across Spanish cities: unveiling the heterogeneity of gentrification

This study identifies and maps distinct neighbourhood-level trajectories of urban change across nine Spanish cities (Madrid, Barcelona, Valencia, Zaragoza, Sevilla, Málaga, Murcia, Palma, and Las Palmas de Gran Canaria), using a two-step unsupervised machine learning approach that combines UMAP dimensionality reduction with HDBSCAN density-based clustering. Drawing on the 2011 and 2021 Population and Housing Censuses and annual rental price data over the same decade, the analysis covers 657 custom-built census groupings. Five recurring neighbourhood profiles emerge, revealing a broad continuum from relative stability to sustained multidimensional change. Stable working-class neighbourhoods remain dominant across much of the urban periphery, while stronger dynamics of residential turnover, socioeconomic upgrading and rent growth are concentrated in central and intermediate areas. These patterns broadly reproduce a centre-to-periphery spatial structure, but their prevalence and spatial reach vary considerably across cities: more globally connected cities display more extensive and differentiated trajectories of change, while smaller and island cities exhibit more context-specific and, in some cases, polarised configurations. These findings contribute new knowledge by revealing heterogeneity in neighbourhood stability and change across cities, providing a comparative framework for identifying distinct trajectories of urban transformation and informing place-sensitive urban planning.