Spatio-Temporal Clustering Characteristics Analysis Of Urban Traffic Status Based On Cloud GIS

The congestion of urban traffic has been a serious challenge to urban development in recent years.Exploring the reasons behind traffic congestion will provide theoretical support for solving these problems.Taking perspective from Geographic Information System,the paper takes Shahekou District,Zhongshan District,Ganjingzi District and Xigang District of Dalian City as research areas,research areas' local urban road network data are sorted and real-time traffic situation data are collected from GaoDe Map service inorder to match the location of congestion events with the local road network by using match algorithm.For steps mentioned above,the method of mining spatio-temporal clustering characteristics of urban traffic is explored based on spatio-temporal density clustering algorithm and the basic framework of Cloud GIS is initiatively explored,thus the overall analysis process is designed as a prototype system based on cloud GIS.On the basis of studying various clustering algorithms,this paper takes the characteristics of traffic congestion data into account,chooses density-based spatio-temporal clustering algorithm ST-DBSCAN as the main research method to explore the basic principle,analysis flow,algorithm design and parameter selection standard of such algorithm.In the process of literature research it is found that the reliability of calculation results can not be guaranteed only by using ST-DBSCAN.Therefore,it is decided to adopt the universal spatio-temporal clustering analysis method as the theoretical basis.The analysis process mainly includes Exploratory Spatio-Temporal Data Analysis,Spatio-Temporal Clustering Calculation and Effectiveness Evaluation of Spatio-Temporal Clustering.In the stage of Exploratory Spatio-Temporal Data Analysis,we mainly discuss the autocorrelation and stability of spatio-temporal data.Through this step,we can preliminarily determine whether the data is clustered in the spatio-temporal range.Data sets with aggregation state are valuable to explore their clustering characteristics The stage of Spatio-Temporal Clustering Calculation is to select suitable spatio-temporal clustering algorithm for the research object.In the evaluation stage of validity of spatio-temporal clustering results,various evaluation indices are used to gauge whether the selected parameters are reasonable or not.Only through a complete set of spatio-temporal clustering analysis process can the credibility of the final spatio-temporal clustering results be guaranteed.Effective spatio-temporal clustering results can provide the correct spatio-temporal characteristics of data,thus ensuring the reliability of spatio-temporal clustering feature analysis.By initially exploring cloud GIS service mode,a prototype system of spatial-temporal clustering analysis for urban traffic conditions is hence designed and developed and the system structure,interface,function and database design are also introduced.The basic GIS function,exploratory spatiotemporal data analysis function,ST-DBSCAN analysis function and spatio-temporal clustering validity evaluation function are all developed respectively realizing spatio-temporal autocorrelation analysis,spatio-temporal stability analysis,ST-DBSCAN calculation and effectiveness evaluation of spatio-temporal clustering,the outcome for each stage are demonstrated in visualization form.