The non-linear relationship between neighborhood environment and residents' happiness: an empirical study from low-income communities in Beijing

It is widely acknowledged that the quality of the neighborhood environment has an impact on residents' sense of happiness. However, the existing literature provides limited empirical evidence of the nonlinear effects of various attributes of the neighborhood environment on residents' happiness. To fill this gap in the knowledge, this paper employs an explainable artificial intelligence approach, namely, the GBDT-SHAP method. We conducted a case study of low-income communities in Beijing to explore the nonlinear and heterogeneous relationships between the neighborhood environment and the happiness of residents during the urban transformation process. The results show that residents in mature affordable housing communities report a significantly higher level of happiness compared to those in old unit communities and public rental housing communities. The subjective environment plays a more crucial role in influencing residents' happiness. Service facility satisfaction and natural environment satisfaction were identified as the most critical variables affecting residents' happiness, with social identification in the social environment having a more pronounced impact. Among the objective characteristics of the built environment, the POI entropy index and subway accessibility are the most important variables influencing residents’ happiness, with the crime rate in the social environment having a more significant effect. Service facility satisfaction and natural environment satisfaction were positively correlated with happiness. The relationship between crime rate and unemployment rate and happiness exhibits an inverted U-shaped pattern. Additionally, we investigated the differential impacts of community environments on various aspects of happiness among local and non-local residents.