Spatiotemporal evolution of urban livability amid China’s urban renewal: genetic algorithm projection pursuit and geographically weighted regression

Against the backdrop of the concurrent advancement of rapid urbanization and large-scale urban renewal, enhancing urban livability has become an important issue in China’s urban development. This study employs a Genetic Algorithm-optimized Projection Pursuit (GAPP) model to assess livability in 125 Chinese cities (2010–2022). Integrating spatial analysis and Geographically Weighted Regression (GWR), we investigate livability’s spatiotemporal evolution and the spatially heterogeneous impacts of urban renewal. Empirical findings reveal three major characteristics. First, urban livability exhibits robust spatial “disequilibrium and lock-in,” with coastal regions forming high-value clusters, while the Northwest and Northeast remain entrenched in cold spots. Second, regional divergence continues to widen, accompanied by a profound structural mismatch characterized by “macro-level convergence versus micro-level polarization.” Third, while urban renewal catalyzes livability improvement, its efficacy follows the law of diminishing marginal returns. Marginal improvements are pronounced in the baseline-depressed Northeast but severely diminish along the infrastructure-saturated southeastern seaboard. These findings validate the spatial non-stationarity of human habitat evolution. We argue that policymakers should abandon “spatially blind” physical expansion paradigms and instead promote context-specific organic renewal and asymmetric resource allocation to bridge regional livability divides.