Spatial Mismatch and Nonlinear Responses Between Street Built Environment Quality and Subjective Well-Being: Evidence from Beijing

Improvements in urban spatial environmental quality do not necessarily lead to corresponding improvements in residents’ subjective well-being, and spatial mismatch may exist between objective environmental provision and residents’ actual experiences. This study focused on the area within Beijing’s Fifth Ring Road and integrated street-view imagery and geotagged Weibo text at the 500 m × 500 m grid-cell scale to construct a spatial matching framework between street built environment quality (ENV) and expressed subjective well-being (expressed SWB), and employed XGBoost–SHAP to analyze the environmental characteristics of mismatch types. A total of 1100 valid grid cells were obtained. Results showed that ENV and expressed SWB were weakly correlated overall (Pearson r = −0.001, p = 0.965; Spearman ρ = −0.007, p = 0.819); the SWB-above-ENV, ENV–SWB matched, and SWB-below-ENV types contained 367, 366, and 367 grid cells, accounting for 33.4%, 33.3%, and 33.4%, respectively. For the SWB-below-ENV type, Green View had the highest Mean |SHAP| (0.378), and key street environmental characteristics exhibited nonlinear changes such as turning points and leveling-off patterns. By placing the objective street environment and publicly expressed emotions within the same spatial framework, this study provides evidence for identifying spatial differences between environmental provision and residents’ experiences.