Deciphering residential satisfaction among different urban renewal approaches: An integration of gradient boosting decision trees and impact asymmetry analysis

Keywords: Residential satisfaction; urban renewal approaches; community cohesion; machine learning; impact-asymmetry analysis; China
Abstract Type: Paper Abstract

Abstract

This study compares the relationship between changes in social and neighborhood attributes and residential satisfaction (RS) under two urban renewal approaches (urban development and urban rehabilitation) in Chongqing, China. This paper applied a mixed method by combining questionnaires and semistructured interviews to examine the variables influencing RS dynamics. The results from the gradient boosting decision tree showed that “community cohesion”, “part of the community”, “increased public space”, and “higher trust in residents” have relatively large impacts in both models. Using impactasymmetry analysis, this study also illustrates the nonlinear effects of social and environmental attributes on RS. Many attributes, such as “part of the community”, “community cohesion” and “more help from residents”, show different patterns of nonlinear impacts between the two urban renewal projects.Based on quantitative results and qualitative evidence, this study offers policy implications for future urban renewal projects.