Designing walkable streets: Exploring nonlinear effects of street-level built environment on older adults’ walking intention

The growing trend of older adults aging in their long-term residential settings has increased the demand for built environment designs that promote mobility, safety, and social engagement. Streets, as accessible and ubiquitous public spaces, play a critical role in shaping older adults’ walking intention and overall well-being, highlighting the importance of age-friendly street design. However, the relationship between the street-level built environment and older adults’ walking intention remains underexplored, especially regarding potential nonlinear effects. This study employed the DeepLabV3 semantic segmentation model to extract street-level built environment features from street view images. Older adults’ walking intention was quantitatively assessed by applying the Microsoft TrueSkill algorithm to convert pairwise image comparisons into ranked scores. Integrating a Gradient Boosting Decision Tree model with SHAP analysis, we examined the nonlinear and interaction effects of street-level built environment features on walking intention. Results showed that: (1) street-level built environment features were stronger determinants of older adults’ walking intention than socioeconomic factors, with green view index exerting the greatest influence; (2) the effects of street-level features exhibited nonlinear patterns, mainly power or exponential functions, with a threshold at GVI = 0.48 beyond which marginal gains diminished; (3) GVI, pedestrian path ratio, and colour richness were positively correlated with older adults’ walking intention, while building-to-street ratio, proportion of walls, and crowd concentration index showed negative correlations; and (4) significant interaction effects were observed among street-level built environment features. This study advances theoretical understanding of how microscale built environment shapes human behavior through nonlinear and synergistic mechanisms, providing actionable insights for designing age-friendly and walkable communities.