Design And Implementation Of Social Network Visualization Analysis System

Social network originated in humanities and sociology, with the rapid development of network, especially the popularity of Web2.0, internet-based communication or social media also have great development, such as E-mail communication, microblog, blog, etc. These communication and relationship between human produced a large amount of data which reflect the various forms of interpersonal activities, these data implied the various social networks, such as E-mail, criminal networks and co-author network, etc.The technology of network recently has great development, it was found that the real social network has some common features, such as the community structure and key roles. Found that the features of social network is valuable to the analysis of criminal organizations or the transmission of public opinion. However, the actual social network is a huge network of annotation, its nodes and edges have rich attribute information, the current widely used visual layout algorithm ignores the individual links and nodes’ attributes, cannot effectively reflect the influence because of social attribute between people and people in social network. In addition, due to the social network is a large-scale data, nowadays most of visual measures cannot effectively show the network’s overall layout. According to the characteristics of temporal network, the main research content of this paper is putting forward a kind of large-scale visual layout algorithm for temporal relation network, based on complex networks theory mining social network structure and key roles, combined with measures for visualization of social network, thus provides users a good interactive analysis tools for social network’s relationship between people, our innovation achievements are as follows:(1) A large-sclae temporal network visual layout method. The layout algorithm think about the time attribute of links between nodes, combined with the time features new force optimization calculation formula is put forward, using the parallel processing technology to carry on the cutting edge, multi-level layout, finally provides overall network visualization, experiments show that compared to the classical layout algorithm our algorithm has better performance in temporal network layout.(2) To develop a social network visualization analysis system named Ginkgo for analysis found that patterns of real social network. The analysis of the Ginkgo framework consists of the following strategies: network generation, network detection, centrality analysis and statistical analysis, network visualization. Based on the above strategy, the system take into modeling of network, community discovery, centrality analysis and network rendering techniques. The system is not only a general platform for the network analysis, and the effect of the platform can effectively verify the algorithm.