Spatial-temporal heterogeneity and built environment nonlinearity in inconsiderate parking of dockless bike-sharing

Abstract

Although previous studies have shed light on the travel behaviour of dockless bike-sharing (DBS) users, little research focused on their inconsiderate parking behavior. Unlike the travelling behavior, the choice of parking location is closely linked to the different built environments surrounding the parking locations. Therefore, to improve the efficiency of governance, it is vital to explore the parking patterns and heterogeneous influences of the built environment on inconsiderate parking and formulate targeted measures. This paper measures the coordinates of prohibited parking areas in the field to identify inconsiderate parking. Based on big data from Mobike DBS and data on the built environment, the paper empirically analyzes the heterogeneous spatiotemporal distribution patterns of inconsiderate parking with clustering and decision trees. The influencing factors of inconsiderate parking and their nonlinear effects are further analyzed using random forest and partial dependence plots (PDP). The results show that there is significant spatiotemporal heterogeneity in inconsiderate parking, in which different clusters reflect various characteristics of the built environment. Furthermore, marginal effect analysis finds that influencing factors such as riding distance, catering service places, lifestyle services, sports and leisure places, and hotels and hostels have a strong effect on inconsiderate parking behavior, and show nonlinear effects with optimal allocation intervals. Therefore, targeted strategies should be carried out in terms of dynamic temporal adjustment, precise spatial layout, differential management according to time and zone, and cause-assisted administration. The paper’s results provide important decision-making support for inconsiderate parking.