Mining And Visualization Of Taxi Trajectory Data

The GPS terminal will periodically send some real-time status information to the taxi dispatch center,such as the longitude,latitude of the taxi,the direction of travel of the taxi,and whether or not passengers are carrying passengers.Over time,taxi companies collected and saved vast amounts of taxi trajectory data.Nowadays,more and more people are investing in the mining of these data,making it one of the research hotspot of current urban traffic big data.The mining,analysis and utilization of taxi trajectory data provide new methods and new ideas for discovering and solving urban problems.Based on the taxi trajectory data of Beijing,this paper excavates the hot spots of passenger carrying,analyses the spatial and temporal distribution characteristics of passenger arrival and departure points,studies the residents' traveling rules in hot spots of passenger carrying,and analyses the trajectories with the same starting and ending points between hot spots.Specifically include:Firstly,taxi trajectory data preprocessing.The trajectory data is imported into Oracle database by using SQL Loader method.By setting rules and methods of trajectory extraction,redundant data,invalid data and drift points in the trajectory are filtered,and the matching algorithm between the trajectory points and the road is designed.The data in the experiment are matched so that each trajectory point has corresponding road information.Second,the presentation and establishment of a passenger model.According to the research goal of this paper,a passenger model is proposed.It is considered that the passenger-passenger-departure process is a complete passenger,and the part of the taxi is removed.Only the part of the taxi operator is reserved..The data organization structure design and formal expression are carried out for a passenger model.Based on this model,the original trajectory data is extracted and sorted,and a passenger model trajectory database for analysis and application targets is established.The excavation laid the foundation.Third,hotspot extraction and visualization based on DBSCAN clustering.Firstly,the taxi drop-off point data is extracted,and the hotspot library is pre-established according to the statistical information on the network.Then,according to the total number of sunrise trips and the number of trips per hour,the total number of sunrise rows is counted.Based on this,DBSCAN clustering algorithm is used.Generate hot spots for different time periods.Thenuclear density clustering analysis in ArcGIS software was used to visualize the two-dimensional thermodynamic map and the three-dimensional thermodynamic map.Fourth,the trajectory analysis and visualization of the same starting and ending points.Firstly,the trajectory with the same starting and ending points is extracted according to a passenger model,and the trajectory simplification algorithm is programmed to simplify the trajectory.Then,the matching distance of the trajectory segment is used as the basis for measuring the similarity of the trajectory,and the similarity of the extracted trajectory is performed.Analysis,the hotspot path from Xizhimen business circle to financial street business circle: Xizhimen West Second Ring Financial Street business circle,and dynamic visualization of the hotspot path between hotspot areas.