Exploring the relationship between the built environment and metro ridership among older adults: Evidence from Dalian, China

Introduction Despite increasing attention to age-friendly metro planning, it remains unclear which station-area built environment variables are associated with older adults’ metro use. Existing studies have rarely examined weekday-weekend differences or intra-day temporal heterogeneity, and conventional models are often limited in capturing nonlinear relationships and interactions among built environment variables. Methods This study uses metro smart card data and built environment data from Dalian, China. Three conventional benchmark models and five ensemble learning models are compared, and the best-performing light gradient boosting machine (LightGBM) model is interpreted using Shapley Additive Explanations (SHAP) to examine the relative importance, nonlinear relationships, and interactions between built environment variables and older adults’ metro ridership across weekdays and weekends, each divided into four daily periods. Results The results identify road density, the number of shopping malls, and the number of chess/card rooms as the dominant variables related to older adults’ metro ridership. Nonlinear relationships indicate that road density and the numbers of bus stops, shopping malls, chess/card rooms, hospitals, and pharmacies are positively associated with metro use during the daytime off-peak period (9:00-15:29), whereas population density, land-use mixture, and the numbers of intersections, markets, fitness centers, and parks/squares are negatively associated with ridership. Interactions are strongest during the daytime off-peak period (9:00-15:29) and morning peak period (6:30-8:59), revealing that older adults’ metro use is related to combined environmental conditions rather than to single factors alone. Conclusions These findings offer practical insights for optimizing age-friendly metro station services and informing future transport planning, particularly in the context of developing more accessible, coordinated, and supportive mobility systems for aging populations.