Statistical Characteristics And Cluster Analysis Of GPS Trajectory Data For Floating Cars

Under the current background of big data,through in-depth analysis and mining of various types of data,we can grasp the behavior patterns of human activities,and then explore the objective laws of human activities,and become an important way to solve urban problems,in which GPS trajectory data of floating vehicles One of the main manifestations of social data in geographic information environment can study the spatial movement law of the group from the perspective of individuals and provide new data support and research ideas.It is an important way to solve urban problems by exploring trajectory data to explore human movement patterns and patterns of activity,and then to explore deeper levels of knowledge.Through the statistical characteristics and cluster analysis of a large number of GPS trajectory data,it is possible to find out the traveler's travel time and space characteristics and patterns in the city,to perceive the urban rhythm to a certain extent,to tap human social activity records,and to understand the hot spots of urban traffic.,Provide reasonable suggestions for city construction and traffic planning.This article will use the GPS trajectory data of floating vehicles for statistical features and cluster analysis.The main content is reflected in the following aspects:(1)Data preprocessing.The original data is 385 million trajectory data of more than 30,000 taxis in Beijing during the week from May 11 to May 17,2015.The data was preprocessed in SQL Server to get 5,000 for experimental use.More than 10,000 valid data.(2)Statistical analysis of trajectory data.Analyze the statistical characteristics of distance,long-cycle time,single track time,and direction of the load-carrying data after processing,and study the probability distribution model of taxi distance and time,and find the distance and single track.The distance attenuation effect and heavy tailing effect of time travel revealed the characteristics of urban residents' travel on working days and weekends,discovered the hidden day rhythm of long cycle time,and statistically calculated the linear average of the trajectory of the passenger segment.Discover its relationship with urban road traffic planning.(3)Cluster analysis of trajectory data.The longest common subsequence(LCSS)of the selected trajectory is used as a measure of similarity,and it is calculated for 5.13(working days)and 5.16(weekends)from 12:00 to 14:00(lunch peak)and 17:00-19:00(night peak)were processed and density-based DBSCAN clustering was used to determine the distribution pattern of traffic flow clusters in space.It was found that most of them were clustered in the central city and spread to a small part of the surrounding urban areas,and they were different in time.Traffic flow clusters were compared and analyzed,and the evolution characteristics of traffic flow clusters were studied.The results showed that the traffic flow at the mid-peak hours was mainly the flow within the central urban area,and the traffic flow at the late peak was mainly the flow of the central city to the surrounding areas,and the weekend was relatively Workday clustering clusters are more sparsely distributed.