The Visual Analysis Of Traffic Data Based On Semantic Extraction

The massive traffic trajectory data recorded by GPS devices imply urban movement patterns under the space-time environment.It provides city-wide coverage and effectively reveals the information of underlying traffic condition and crowd flow direction.However,the trajectory data are usually characterized by large scales,high complexities and multiple dimensions.Besides,there are the problems such as low sampling rates and large errors related with the traffic data.It is a challenging yet meaningful research to effectively analyze the massive complex traffic data and find the useful information hidden in the data.In this paper we study the GPS trajectory data with data mining and visualization techniques.Through providing the interactive visual analysis of the underlying traffic patterns of the city,the work of this paper is applied to guide the users to choose proper area for stores who want to set up shops to do business.The main work is as follows:(1)Traffic data analysis based on the LDA topic model: In this paper we construct the LDA topic model to analyze the GPS taxi trajectory data.The topic information are combined with the traffic volume information to choose the representative candidate areas.(2)Traffic flow visualization based on semantic analysis: We generate traffic flow graphs between candidate areas to help users understand the distribution of the candidate areas and the taxi running rules.We study the distribution and semantics of the topics from three aspects: time,space and point of interest.The visualization and analysis scheme of the LDA topics analysis is proposed to help users understand the urban areas' characteristics better by combining with topic semantics.In addition,the users are enabled to explore the correlation of trajectories and each topic to support the detailed trajectory analysis.(3)The visual exploration of candidate areas: The attribute data which measure the traffic condition are extracted from the traffic data.Inspired by the wheel of vehicles,we design a metaphor-based glyph to summarize the multi-dimensional attribute features of each candidate area.Users can explore the general traffic attributes over time of interested area through varied interactions to learn the details of the area from multiple perspectives.(4)The visual analysis prototype system of traffic trajectory data: In order to provide guidance for potential business to select proper sites for their shops,we design and implement the visual analysis prototype system of traffic trajectory data.The system provides three main views to visualize the LDA topics and candidate area attributes.Besides,a variety of user interactive operations are supported to allow users effectively explore the system.Finally,the feasibility and validity of the system is verified by a case study.