Identifying the natural clusters of nodes in a graph and treating them as supernodes or metanodes for a higher level graph (or an abstract graph) is a technique used for the reduction of visual complexity of graphs with a large number of nodes. In this paper we report on the implementation of a clustering algorithm based on the idea of ratio cut, a well known technique used for circuit partitioning in the VLSI domain. The algorithm is implemented in WINDOWS95/NT environment. The performance of the clustering algorithm on some large graphs obtained from the archives of Bell Laboratories is presented.

Two tasks in Graph Visualization require partitioning: the assignment of visual attributes and divisive clustering. Often, we would like to assign a color or other visual attributes to a node or edge that indicates an associated value. In an application involving divisive clustering, we would like to partition the graph into subsets of graph elements based on metric values in such a way that all subsets are evenly populated. Assuming a uniform distribution of metric values during either partitioning or coloring can have undesired effects such as empty clusters or only one level of emphasis for the entire graph. Probability density functions derived from statistics about a metric can help systems succeed at these tasks. CR Categories and Subject Descriptors: I.3.6 [Computer Graphics]: Methodology and Techniques -- Interaction Techniques; I.3.8 [Computer Graphics]: Applications Additional Keywords: graph visualization, graph navigation, metrics, clustering 1. INTRODUCTION A key issue...

GraphXML is a graph description language in XML that can be used as an interchange format for graph drawing and visualization packages. The generality and rich features of XML make it possible to define an interchange format that not only supports the pure, mathematical description of a graph, but also the needs of information visualization applications that use graph--based data structures. 1999 ACM Computing Classification System: D.2.12, H.3.5, I.3.6, I.3.8, I.7.2, Keywords and Phrases: information visualization, graph visualization, user interfaces, XML Note: The work was carried out under the project INS3.1 "Information Visualization". 1 INTRODUCTION GraphXML is a graph description language in XML * . The goal of GraphXML is to provide a general interchange format for graph drawing and visualization systems, and to connect those systems to other applications. The requirements of information visualization have greatly influenced design decisions during the development of Grap...

Planar drawings of clustered graphs are considered. We introduce the notion of completely connected clustered graphs, i.e. hierarchically clustered graphs that have the property that not only every cluster but also each complement of a cluster induces a connected subgraph. As a main result, we prove that a completely connected clustered graph is c-planar if and only if the underlying graph is planar. Further, we investigate the influence of the root of the inclusion tree to the choice of the outer face of the underlying graph and vice versa.

Identifying the natural clusters of nodes in a graph and treating them as supernodes or metanodes for a higher level graph (or an abstract graph) is a technique used for the reduction of visual complexity of graphs with a large number of nodes. In this paper we report on the implementation of a clustering algorithm based on the idea of distance-k cliques, a generalization of the idea of the cliques in graphs. The performance of the clustering algorithm on some large graphs obtained from the archives of Bell Laboratories is presented.

Several previous systems allow users to interactively explore a large input graph through cuts of a superimposed hierarchy. This hierarchy is often created using clustering algorithms or topological features present in the graph. However, many graphs have domain-specific attributes associated with the nodes and edges which could be used to create many possible hierarchies providing unique views of the input graph. GrouseFlocks is a system for the exploration of this graph hierarchy space. By allowing users to see several different possible hierarchies on the same graph, the system helps users investigate graph hierarchy space instead of a single, fixed hierarchy. GrouseFlocks provides a simple set of operations so that users can create and modify their graph hierarchies based on selections. These selections can be made manually or based on patterns in the attribute data provided with the graph. It provides feedback to the user within seconds, allowing interactive exploration of this space.

Abstract. Compound-fisheye views are introduced as a method for the display and interaction with large graphs. The method relies on a hierarchical clustering of the graph, and a generalization of the traditional fisheye view, together with a treemap representation of the cluster tree. 1

Abstract. In this paper we present a novel approach for cluster-based drawing of large planar graphs that maintains planarity. Our technique works for arbitrary planar graphs and produces a clustering which satisfies the conditions for compound-planarity (c-planarity). Using the clustering, we obtain a representation of the graph as a collection of O(log n) layers, where each succeeding layer represents the graph in an increasing level of detail. At the same time, the difference between two graphs on neighboring layers of the hierarchy is small, thus preserving the viewer's mental map. The overall running time of the algorithm is O(n log n), where n is the number of vertices of graph G.

This paper presents a technique for visualizing the differences between two graphs. The technique assumes that a unique labeling of the nodes for each graph is available, where if a pair of labels match, they correspond to the same node in both graphs. Such labeling often exists in many application areas: IP addresses in computer networks, namespaces, class names, and function names in software engineering, to name a few. As many areas of the graph may be the same in both graphs, we visualize large areas of difference through a graph hierarchy. We introduce a path-preserving coarsening technique for degree one nodes of the same classification. We also introduce a path-preserving coarsening technique based on betweenness centrality that is able to illustrate major differences between two graphs.

We propose a method for visualizing a set of related metabolic pathways across organisms using 2 1/2 dimensional graph visualization. Interdependent, twodimensional layouts of each pathway are stacked on top of each other so that biologists get a full picture of subtle and significant differences among the pathways. The (dis)similarities between pathways are expressed by the Hamming distances of the underlying graphs which are used to compute a stacking order for the pathways. Layouts are determined by a global layout of the union of all pathway graphs using a variant of the proven Sugiyama approach for layered graph drawing. Our variant layout approach allows edges to cross if they appear in different graphs.