Upload
liangou
View
612
Download
0
Tags:
Embed Size (px)
DESCRIPTION
Network data often contain important attributes from various dimensions such as social affiliations and areas of expertise in a social network. If such attributes exhibit a tree structure, visualizing a compound graph consisting of tree and network structures becomes complicated. How to visually reveal patterns of a network over a tree has not been fully studied. In this paper, we propose a compound graph model, TreeNet, to support visualization and analysis of a network at multiple levels of aggregation over a tree. We also present a visualization design, TreeNetViz, to offer the multiscale and cross-scale exploration and interaction of a TreeNet graph. TreeNetViz uses a Radial, Space-Filling (RSF) visualization to represent the tree structure, a circle layout with novel optimization to show aggregated networks derived from TreeNet, and an edge bundling technique to reduce visual complexity. Our circular layout algorithm reduces both total edge-crossings and edge length and also considers hierarchical structure constraints and edge weight in a TreeNet graph. These experiments illustrate that the algorithm can reduce visual cluttering in TreeNet graphs. Our case study also shows that TreeNetViz has the potential to support the analysis of a compound graph by revealing multiscale and cross-scale network patterns.
Citation preview
TreeNetViz: Revealing Patterns of Networks over Tree Structures
Liang Gou and Xiaolong (Luke) Zhang
The Penn State University
Introduction
• A compound graph with a network and a tree is seen in many domains.• A scientific collaboration network with author affiliation
information
The Problem
• How to visualize and explore the patterns of a network over a tree?• Mutual influence between two sub-graphs
• Relationship among entities at different levels
• Multiscale relationship:
• Cross-scale relationship:
Network Aggregation
• We need to aggregate a network over trees.
• An aggregated network can be generated by
• applying a cut on a tree to specify nodes of interest; • aggregating edges based on the nodes in a cut.
Our Goal in TreeNetViz
• Interactive visualization to • Provide representations of both tree and
network structures with an integrated view;
• Support interactive aggregation of a network over tree structures;
• Allow exploration of the patterns of aggregated networks in both multiscale and cross-scale style.
Tree as Radial Space Filling (RSF) Layout
• A RSF technique to show a tree structure
• An aggregated network can be placed over
a RSF tree
Network as Circular Layout
• TreeNetViz uses a circular layout to show node connections in a network
A circular layout of a network overlaid on RSF tree
A RSF circular layout of an aggregated network
Hierarchical Edge Bundling (Holten, 2006)
Interactions
• Multiscale view: users can control the scale level.
• Cross-scale view: users can expand or collapse a node.
Interactions (Cont.)
• An ego-network view and a critical path view
An ego-network view A critical path view
Algorithm to Reduce Visual Complexity:Hierarchy-awareness Weighted Circular Layout
• Hierarchy-awareness Weighted Circular Layout (HWCL) :• Sort nodes along a circle to
reduce visual complexity• Consider the restriction of a
tree structure• Child nodes are sorted
under their parent node• Consider the weight of edge
The Key Idea
• Optimize node positions to achieve• Fewer (weighted) edge crossings• Less total (weighted) edge length
• A heuristic approach tries different combinations of node order by switching node pairs
Switch to reduce edge crossings Switch to reduce total edge length
Visual Improvement By HWCL
Without HWCL With HWCL
Case Study: A Co-authoring Social Network
• Help us understand collaboration patterns among diabetes researchers across a research university M.
The dataset847 authors (nodes)2,498 co-author links.
Social affiliation hierarchy10 colleges/schools90 departments/centers
Multiscale Exploration: Collaboration among Colleges
Multiscale Exploration: Collaboration among Departments
Multiscale Exploration: Collaboration among Individual Researchers
Cross-Scale Exploration:Collaboration between Departments in a College and other Colleges
Cross-Scale Exploration:Collaboration between Individuals in a Department with other Departments and Colleges
Cross-Scale Exploration: Find a Potential Research Broker between Two Organizations
Future Work
• Enable users to modify a tree structure• Compare TreeNetViz with other techniques
• Node-link diagrams• Matrix view
(ASK-GraphView, 2006) (Honeycomb, 2009)
TreeNetVizVideo & Application: zhang.ist.psu.edu/leon
Liang Gouemail: [email protected] Poster Session @6:15pm, EastPoster Area