Research On Large-scale Dynamic Graph Visualization

Graph data,such as social network data appears frequently in our life.At the same time,it is increasingly important to study and analyze the characteristics of such graph data.As the most important part of using visual method to study graph data,graph layout algorithm directly determines the quality of visual analysis.Considering that the current graph data analysis has three characteristics: large-scale,timeliness and uncertainty,this paper focuses on the large-scale dynamic graph layout algorithm.Firstly,to solve the problem that the existing dynamic graph layout method is difficult to maintain dynamic information and ensure the balance of the layout aesthetics,moreover,the algorithm is complex and hard to accelerate,this paper proposes an heuristic dynamic graph layout algorithm to introduce the idea of inverse Markov analysis to maintain the beauty of the graph layout while preserving the mental map as much as possible.Because of the increasing amount of data,the simple layout of small dynamic graph can not meet the needs of visual analysis because of the efficiency of operation and other reasons,so this paper firstly studies and optimizes the existing static hierarchical layout algorithm,then transforms the largescale static hierarchical layout into large-scale dynamic layout by solving the problem of dynamic hierarchical layout.In this paper,we firstly compared our dynamic graph layout algorithm and large-scale dynamic layout algorithm with the existing algorithm by experiment.The experiment verifies the effectiveness of our algorithm from two aspects of data and view.Because the visualization problem is more about the user's visual perception than the data.For each observer,their use experience is different.Therefore,we designed a user survey at the same time to judge the efficiency and quality of our algorithm by collecting the user's perception and feedback.Finally,we design a simple graph visualization analysis system based on large-scale dynamic graph layout algorithm to show the effect of graph layout and graph visualization on graph information analysis.