Visual Analytics Of Road Traffic With Large Scale Taxi Gps Data

Due to its intuitive performance to the hidden rules of abstract data, Information Visualization is gradually being the critical technique for data analysis. With the development of the urban road monitoring equipment, GPS and widespread of mobile handheld devices, as well as massive stored data, huge amounts of geographical data need analyzing efficiently and properly. Now how to dig the valuable information from the massive data has been a headache but challenging issue. However, the rise of Intelligent Transportation System(ITS) promotes the massive traffic data research in recent years. Traffic data visualization is an important part of ITS aims at visualizing big traffic data. Confronted with the trend of digital life and the big data challenge, classic visual chart may not perform well when dealing with complicated data, e.g. pie chart, histogram and line chart clearly visualize small data set but would be disordered to show big data. Traffic data visualization can solve this problem.Embedded visual component is a innovation point in this paper. Visual components are equipped with excellent flexibility that could accurately show the dataset with large data volume and large time span. And seam carving algorithm is taken to broaden the narrow roads in map into specific width without twisting the perimeter zone of roads, with the novel arrow graph and stacked graph in them. The paper explore the city traffic condition in micro perspective. The main work and achievements are as follows:(1) Analysis method with embedded map. General spatio-temporal data visualization is based on dynamic map that combines with other visual multi-window. Though it can help users to understand the data, it burdens their difficulty in reading charts, and causes confusions. Mapping data to simple points or line in the map is also utilized in existing visualization. However, this paper broadens the narrow roads in map road to accommodate the visual components to analyze moving target rules in space with the road context.(2) Traffic flow analysis based on road section and intersection. Currently traffic visualization applications at home and abroad are mostly geared to macroscopic traffic problems, such as the needs of the global urban traffic condition, or whether the roads are crowded or not and etc. Traffic on the road section and intersection is the core part in this paper. For example, observing driving rules on the road and abnormal behavior from the micro perspective to solve the existing traffic problems.(3) Design for embedded visual components. On the issues like blocked visual elements and inadequate room on the map that appears in using of visual technique, novel arrow graph and stacked graph are proposed to the traffic situation analysis which have clear visual effect to the big data.The preprocessed taxi GPS data are stored in a cloud computing platform in a distributed manner. With the MapReduce to accelerate data inquiry and processing, the data storing and handling is given to clouds platform, and the data after computation is displayed in front-end. Arrow graph and stacked graph in road section mentioned before can directly reflect rules and find out the temporal relations. Arrows at intersections can be used to judge the rationality of traffic light, and relationship of cars in section and intersection in all directions by observing the car flow, especially to find out the reasons cause traffic jams.