Research on spatiotemporal heterogeneity and flow characteristics of dockless bike-sharing big data in riding demand

Inhalt:
Dockless bike-sharing system brings a new urban transportation mode, and users can expediently start and end riding in areas that do not exceed the stopping range. Analyzing the spatiotemporal characteristics in its usage through bike-sharing big data can provide fine-grained insights into urban dynamics. In this study, from the perspective of singular value decomposition and complex network, the spatio-temporal heterogeneity and flow characteristics of riding demand in dockless bike-sharing under the normalization of epidemic prevention and control are explored. Research shows that: 1) There are various riding demand patterns in the use of shared bicycles. The space-time matrix can be well decomposed through SVD, and the spatial and temporal heterogeneity of the demand patterns can be visualized significantly. Among them, the three with the largest proportion are daily basic demand, strong morning peak demand and evening peak demand. 2) Shared bicycles show multiple hotspots in cycling network space, the use of shared bicycles is mainly concentrated in the west of Shenzhen, forming an ellipse that is rotated 45deg counterclockwise. 3) Combo algorithm can divide the cycling network to an excellent degree, maximize the connection within the community and reduce the connection between the communities. After lots of iterations, 17 communities are detected, and the modularity reaches 0.831. Areas with large clustering coefficients are more likely to form community centers, several adjacent administrative streets form the same community, and the boundary of the administrative street forms the community boundary.