Graph visualization techniques for conceptual Web site traffic analysis

Abstract

Systems that support Web site traffic analysis are core business intelligence applications for many companies. Recent papers remark that these systems are especially useful if they measure the users' interest into the relevant concepts described in a Web site rather than counting users' accesses to the distinct pages forming the Web site. This paper extends existing measures of conceptual Web site traffic analysis and describes a system, called COWA, that supports this analysis by means of network models and graph visualization technologies. The graph drawing algorithmic core of the user interface of COWA is a force directed heuristic that computes a simultaneous embedding of two non-planar graphs. This heuristic optimizes the visualizations in terms of crossing resolution and user's geodesic tendency. Experimental results and case studies show the effectiveness of the proposed approach in practice.

A Heuristic Approach Towards Drawings of Graphs with High Crossing Resolution

“…Most of the approaches that could be useful in practice are based on force-directed techniques [12,19]. COWA is a system that supports conceptual web site traffic analysis [16]; its algorithmic core is a force-directed heuristic to compute simultaneous embeddings of two non-planar graphs with high crossing resolution. Didimo et al [17] describe topology-driven force-directed heuristics to achieve good trade-offs in terms of number of edge crossings, crossing resolution, and geodesic edge tendency; the obtained drawings, however, are not straight-line.…”

Visualizing Group Structures in Graphs: A Survey

“…As an alternative to color, the style of the links within node-link diagrams can be varied to visualize which group an edge belongs to [JJ10,Kit03,SGKS15] , e.g. using solid and different types of dashed links (Figure 8a We identified three juxtaposed visualization techniques [BPD11, DLR10,RMM15], which can all be used for overlapping groups . All three techniques show the graph topology using a node-link diagram .…”

“…We identified three juxtaposed visualization techniques [BPD11, DLR10, RMM15], which can all be used for overlapping groups . All three techniques show the graph topology using a node‐link diagram .…”

Topology-Driven Force-Directed Algorithms

“…Let e 1 , e 2 be two crossing edges of Γ 2 that form an angle smaller than α. In order to enlarge the crossing angle formed by e 1 = (u 1 , v 1 ) and e 2 = (u 2 , v 2 ) we adopt a technique similar to that described in [5]. Denote by c the crossing point of e 1 , e 2 .…”

“…Only a few examples of force directed algorithms use edges with bends (see, e.g., [5,10,12]); the vast majority of the force-directed algorithms represent the edges as straight-line segments and thus the computed drawings are optimal in terms of geodesic edge tendency. Unfortunately, force directed algorithms may introduce unnecessarily many crossings.…”