Methods for Software Visualization of Large Graph Data Structures

In this paper, three different methods for software visualization of large graph structures, respectively Rectangle, Intersection and Combined are presented. The basic concepts for using software development environments are outlined. Their capabilities for visual designing and event-oriented programming are discussed. A brief analysis of the basic features of the environment used to develop the ClipRect Monitor application is made. The main functions of this software are also presented. All experimental results in this study are generated with this application. According to the methodology, six graphs are prepared to determine the effectiveness of the three methods. The number of vertices and the edges of these graphs are proportional to the size of the drawing area (canvas). The drawing areas are also six and have different sizes, such that each subsequent area has a height and width twice the size of the previous one. Besides, for all areas, the width/height ratio is exactly 16:9. This ratio is widely used in monitors as well as laptops, mobile phones and tablets. The largest drawing area that the ClipRect Monitor application scanned during the experiments is 128 000 x 72 000 pixels. This scan is performed for graph G_6 with 1 415 vertices and 100 000 edges. The visualization area is diagonally positioned relative to the drawing area. For each visualization area, each of the three methods, respectively Rectangle, Intersection and Combined is performed. The Combined method executes the Rectangle method first and then the Intersection method. The results show that the Intersection method was the slowest compared to the other two methods in terms of the number of edges of the graph that are analyzed. When the visualization area is internal to the drawing area, the Rectangle method performs better than the Combined method. The Rectangle method gives the best result in terms of time for analysis and drawing of the edges of the graph. The Combined method combines the characteristics of the other two methods. This method is optimal in terms of the time of analysis of the need to draw the edges of the graph relative to the number of drawn edges.