Research On Visual Analytics Of Traffic Big Data

The era of big data,data is growing at Zb level,but the data storage,management,analysis and application of the technology is relatively backward.So in academic circles,it has become the major challenge to extract useful information from massive data.Because visual analytics include the complete process of information acquisition,data processing,human-computer interaction,knowledge representation,reasoning,decision support and collaborative analysis of the exchange of ideas,becoming an ideal tool to achieve this goal.Therefore,based on the analysis of traffic data,with a sense coil flow data and taxi GPS trajectory data based on traffic conditions.In this paper using the clustering algorithm,visual analysis method study the traffic situation of Zhejiang Province of Jiashan County.Jiashan county area to find easy road congestion,hot spots,so as to assist the traffic management department of decision.To sum up,the main works of this paper include:(1)From the wide application of intelligent transportation,the impact of congestion on the city,and the necessity of visual analysis elaborated the paper research background and significance.It summary traffic data preprocessing,data visualization,hot spots,visual analysis of research status at home and abroad,and understand the problems in present research.(2)According to the different methods of data collection of traffic data classification,detailed description on the characteristics of every type of traffic data.It introduced ground sense coil traffic data and the theory of taxi GPS trajectory data preprocessing method,and the data cleaning.And compared with the original data,determined the effectiveness of the pretreatment method.(3)Traffic data in the sense of coil and taxi GPS trajectory data,on the basis of detailed description.Ground sense coil traffic data by means of visual intuitive analysis of jiashan county traffic.Using the vehicles mean speed to calculate the average travel speed,judging the road congestion level with the control section of traffic state classification table.And have found congestion patterns in the spatial scale and time scale.Two ways to complement each other,traffic situation of jiashan county were analyzed.(4)By the summary and analysis of the basic concepts and research status of hot spots,it is found that the lack of the existing hot spots.The clustering algorithm is described in detail,the clustering algorithm can be divided into partition method,hierarchical method,density method,grid method,and the model method.This paper,Single connection hierarchical clustering method is used for optimization.The hot region is more effective by clustering algorithm and has a guiding significance to the traffic management of traffic.