Analysis of Urban Mass Crowd Traveling Patterns Based on Mobile Phone Navigation Trajectory Data

Abstract:

Mobile phone navigation trajectory data has a variety of traffic modes， reflecting the activities of mass crowd， which is suitable for the study of traveling patterns in different traffic modes. Based on mobile phone navigation data， the LightGBM model for traffic mode classification is first proposed to obtain the trajectories of the population in three transportation modes： walking， motorized， and non-motorized mode. Based on these three types of modes， the analysis indexes of time， space， and distance traveling patterns on weekends and weekdays are given， and an experimental analysis is conducted with the mobile phone navigation data of millions of people in Shanghai for four days. The results show that in terms of time distribution， the weekend traveling peak of the residents in Shanghai is later and shorter than that of weekdays， and traveling modes are mainly motor vehicles and walking. In terms of spatial distribution， motor vehicles are mainly concentrated in elevated areas， walking is mainly concentrated near subway stations， the guidance signs of elevated roads and subway stations are not sufficient， and there are more traffic hubs and shopping hotspots on weekends than on weekdays. In terms of distance distribution， navigation traveling distance conforms to the power-law distribution of intercepted segments， and the navigation traveling of the crowd is dominated by short and medium distances， and decays rapidly with distance growth. The research results can provide theoretical basis and technical support for urban planning and urban traffic management policy formulation.

Key words:mobile phone navigation trajectory data;transportation mode recognition;traveling patterns