DiffusionInsighter: Visual Analysis of Traffic Diffusion Flow Patterns

Traffic jam has become a severe urban problem to most metropolises in the world. How to understand and resolve these traffic problems has become a global issue. In the new era of big data, visualization and analysis with traffic-related data are increasingly appreciated. This paper presents DiffusionInsighter, a web-based visual traffic analysis system, that allows users to explore the traffic flow and diffusion patterns with different spatial and temporal granularity. The DiffusionInsighter first applies a visual data cleaning and filtering component to remove dirty data and remain available ones for further analysis. A set of carefully designed interaction and visualization tools including geographical view, pixel map view, chord diagram and network diffusion view is proposed in the DiffusionInsighter to support level-of-detail exploration of diffusion patterns of the traffic flow. Different views are collaborated together and are integrated into geographic map. A series of real-life case studies are conducted using a large GPS trajectory dataset of taxis in Hangzhou.