Research On Urban Residents' Travel Hotspots And Travel Modes Based On Trajectory Data

With the progress of information collection technology,large sample,multi-dimensional and fine-grained GNSS(Global Navigation Satellite System)trajectory data are more and more easy to be obtained.However,due to the lack of effective multi-dimensional data analysis and mining means,it is difficult to effectively mine and use massive resident travel information.It has become an important way to optimize the urban internal spatial structure by identifying traffic hot spots and mining the rules of urban residents' activities and the characteristics of residents' travel patterns by using residents' travel trajectory data.For this reason,trajectory data mining has also become an emerging research hotspot in GIS(Geographic Information Science)and other related disciplines.This paper takes the downtown area of Haikou as the research area and uses multiple source data such as the travel trajectory data of residents.From the perspective of residents' travel hot spots and travel patterns,on the one hand,the characteristics of residents' travel time and space are analyzed to identify residents' travel hot spots and analyze the space-time interaction rules between hot spots.On the other hand,the characteristics of residents' travel patterns and the changes of functional characteristics of travel areas under different dimensions are analyzed.The main research contents and conclusions of this paper are as follows:(1)Analysis of residents' travel time and space characteristics and identification of travel hotspots.First,extract the OD(Origin/Destination)data of residents' travel to study the characteristics of residents' travel time and space.It is found that there are 5 peak time periods for residents' travel on working days,and 4 peak time periods for residents' travel on rest days,and residents are in the travel frequency on weekdays is higher than that on rest days.A plane core density analysis of the OD data of 9 peak travel time periods reveals that the activity of residents' travel increases from four weeks to the center,and there is a phenomenon of multiple centers in residents' travel.Then based on the method of plane kernel density estimation,to establish a model to identify hotspots,to identify hotspots,to study the distribution of hotspots,and to analyze the trend of changes in hotspots.The results show that this model has high recognition accuracy.There are some continuous hotspots in densely populated areas,and the heat value of the continuous hotspots remains at a high level for a long time,reflecting the characteristics of agglomeration of residents' travel activities.(2)Spatial interaction analysis in hot spots for residents' travel.Combining with the theory of complex networks,construct a spatial interaction network for residents' travel hotspots,and analyze the spatial interaction characteristics between hotspots.Through experiments,it is found that the interaction strength between hotspots is attenuated by distance.Only a few hotspots such as popular business districts and large transportation hubs have strong spatial interactions.Most of the hotspots have weaker spatial interactions,and hotspot areas The interactive network also has the characteristics of small world and scale-free.(3)Resident travel patterns and correlation analysis in time and space dimensions.Based on the principle of tensor decomposition,a third-order resident travel starting point tensor and end point tensor are constructed,and the laws of resident travel patterns of different dimensions are explored.The results show that there are two modes in the date dimension,namely working day and rest day mode;there are four modes in the time period dimension,namely morning peak,daytime,evening peak and night mode;in terms of space dimension,the six travel modes obtained will be the study area is divided into different groups.(4)Analyze the functional characteristics of residents' travel areas.According to the results of tensor decomposition,the spatial-temporal interaction characteristics of residents' travel patterns are analyzed,and the functional characteristics of the travel area are inferred.Finally,combined with the POI(Point of Interest)data of the study area,the functional properties of the travel area are reasonably determined.The results show that the travel areas in different time periods correspond to the residence,Commerce,transportation,leisure and entertainment,and other functional characteristics,and the functional characteristics of the travel area will dynamically change over time,and the analysis results are basically consistent with actual cognition.