70 likes | 217 Views
Massive Graph Visualization System. Lei Wang. Introduction. Motivation Graphs are everywhere Seeing is believing Goal An interactive massive graph visualization system. Courtesy of cytoscape.org . Basic Structure: Hierarchy Tree. Two constrains IH and IL Server-side Client-side
E N D
Massive Graph Visualization System Lei Wang
Introduction • Motivation • Graphs are everywhere • Seeing is believing • Goal • An interactive massive graph visualization system Courtesy of cytoscape.org
Basic Structure: Hierarchy Tree • Two constrains • IH and IL • Server-side • Client-side • Children nodes form a graph partition
Server-side Algorithm • Divide-and-conquer: • Recursively break a graph by removing unimportant edges until it is small enough • Operate on hard disk
Client-side Algorithm • Interactively process a subgraph on client side • Finish a branch of hierarchy tree • Preprocessing • Peel: Iteratively remove degree one node • Biconnected component node decomposition
Visualization • Node-link layout • Force directed • Important nodes are attracted to center • Nodes repulse with each other • 3 Dimensional layout, users could change views • Click node to expand\collapse
Experimental Results loc-Brightkite, 58,228 nodes and 214,078 edges Enron emails, 36,692 nodes and 36,7662 edges