The edge, which can encode relational data in graphs and multidimensional data in parallel coordinates plots, is an important visual primitive for encoding data in information visualization research. However, when dat...The edge, which can encode relational data in graphs and multidimensional data in parallel coordinates plots, is an important visual primitive for encoding data in information visualization research. However, when data become very large, visualizations often suffer from visual clutter as thousands of edges can easily overwhelm the display and obscure underlying patterns. Many edge-bundling techniques have been proposed to reduce visual clutter in visualizations. In this survey, we briefly introduce the visual-clutter problem in visualizations. Thereafter, we review the cost-based, geometry-based, and image-based edge-bundling methods for graphs, parallel coordinates, and flow maps. We then describe the various visualization applications that use edge-bundling techniques and discuss the evaluation studies concerning the effectiveness of edge-bundling methods. An edge-bundling taxonomy is proposed at the end of this survey.展开更多
The node-link diagram is an intuitive way to depict a graph and present relationships between entities. Addressing the visual clutter induced by edge crossing and node-edge overlapping is a challenging task as the siz...The node-link diagram is an intuitive way to depict a graph and present relationships between entities. Addressing the visual clutter induced by edge crossing and node-edge overlapping is a challenging task as the size of graph outgrows the visualization space. Many edge bundling methods are proposed to disclose high-level edge patterns. Though previous methods can successfully reveal the skeleton graph structure, the relation patterns at the individual node level can be overlooked. In addition, most edge bundling algorithms are computationally complex, which prevents them from scaling up for extremely large graphs. In this article, we extend SideKnot, an efficient edge bundling method to cluster and knot edges at the node side. Our proposed method is light, runs faster than most existing algorithms, and can reveal the relation patterns at the individual node level. Our results show that SideKnot can disclose a node's standing in the graph as well as the directional connection patterns to its peers.展开更多
基金supported by Foundation for Distinguished Young Talents in Higher Education of Guangdong, China (No. LYM11113)the National Natural Science Foundation of China (Nos. 61103055 and 61170204, and 61232012)
文摘The edge, which can encode relational data in graphs and multidimensional data in parallel coordinates plots, is an important visual primitive for encoding data in information visualization research. However, when data become very large, visualizations often suffer from visual clutter as thousands of edges can easily overwhelm the display and obscure underlying patterns. Many edge-bundling techniques have been proposed to reduce visual clutter in visualizations. In this survey, we briefly introduce the visual-clutter problem in visualizations. Thereafter, we review the cost-based, geometry-based, and image-based edge-bundling methods for graphs, parallel coordinates, and flow maps. We then describe the various visualization applications that use edge-bundling techniques and discuss the evaluation studies concerning the effectiveness of edge-bundling methods. An edge-bundling taxonomy is proposed at the end of this survey.
基金Supported by the National High-Tech Research and Development (863) Program of China (No.2012AA120903)the National Natural Science Foundation of China (No. 61003193)+1 种基金Commonweal Project of Science and Technology Department of Zhejiang Province(No. 2011C21058)the China Postdoctoral Science Foundation (No. 2011M501004)
文摘The node-link diagram is an intuitive way to depict a graph and present relationships between entities. Addressing the visual clutter induced by edge crossing and node-edge overlapping is a challenging task as the size of graph outgrows the visualization space. Many edge bundling methods are proposed to disclose high-level edge patterns. Though previous methods can successfully reveal the skeleton graph structure, the relation patterns at the individual node level can be overlooked. In addition, most edge bundling algorithms are computationally complex, which prevents them from scaling up for extremely large graphs. In this article, we extend SideKnot, an efficient edge bundling method to cluster and knot edges at the node side. Our proposed method is light, runs faster than most existing algorithms, and can reveal the relation patterns at the individual node level. Our results show that SideKnot can disclose a node's standing in the graph as well as the directional connection patterns to its peers.