Discovery And Visual Analytics Of Spatio-temporal Patterns Of Bus Passenger Flocks

The public transportation system plays an important role in modern cities.It provides the public with important shared and large-scale transportation services.However,due to their increasing complexity,it is extremely challenging to mine rules from massive data and to visually analyze and explore them.Passengers who travel by public transportation follow a variety of patterns when traveling,among which the saliency travel pattern is of great significance to understanding the hotspot time and space travel needs of the public transportation system.For public transportation planning managers,planning and management measures such as the location of new lines,the control of departure frequency,the determination of the time of the first and last trains,and the command and dispatch during the operation of the line must be based on a significant spatiotemporal model that fully understands the travel needs of users.For bus passengers,understanding the temporal and spatial laws that are prone to congestion can better combine their own actual conditions to effectively cross-peak travel and improve the comfort of bus travel.In particular,in the special period of high incidence of infectious diseases,how to avoid peak crowds and reduce their own risk of infection is an important aspect that passengers need to consider when traveling.This article refers to the flock definition of unconstrained trajectory to define a flock spatio-temporal pattern in the public transportation system,and the flock pattern mining algorithm including original flock extraction and fusion of the original flock is given.The algorithm consists of five parts,namely,travel chain recovery,establishment of spatio-temporal index,establishment of spatio-temporal matrix,original flock extraction and fusion of original flocks.This paper proposes five visual analysis tasks for public transportation flocks,namely,visualizing the flock,exploring the source and destination of flock members,visualizing the principle of flock mining,autonomously adjusting parameters and mining flocks,and exploring the temporal and spatial patterns of flock.In order to complete the above visual analysis tasks,this article designed a series of flock visual analysis tools and developed a flock visual analysis prototype system.This article designed the particle animation on the map view to represent the flock;the integrated expression of the flock under the polar coordinate heat map framework is designed to discover the tidal flock;the OD import and flow diagram of the flock is designed to help exploring the source and destination of flock members;the Cartesian coordinate heat map of the flock is designed to visualize the two-dimensional space-time matrix of the flock occurrence;the form control part of the flock mining,the line graph of the daily change of the flock number and the flock time travel module are designed to help users dig into flocks independently and explore the temporal and spatial patterns of flocks.Finally,the effectiveness of the system is verified by the visual analysis results of real Shenzhen public transportation data.Based on the bus flocks extracted in this article,passengers can travel on a staggered peak to improve travel comfort;traffic planners can plan and dispatch hot-spot travel sections.