Readings Covered Further Readings Hermann Survey Graph - PowerPoint PPT Presentation

Readings Covered Further Readings Hermann Survey Graph Visualisation in Information Visualisation: a Survey. Ivan Animated Exploration of Graphs with Radial Layout. Ka-Ping Yee, Danyel true survey, won’t try to summarize here! Fisher, Rachna Dhamija, and Marti Hearst, Proc InfoVis 2001, p 43-50. Herman, Guy Melancon, M. Scott Marshall. IEEE Transactions on Visualization and Computer Graphics, 6(1):24-44, 2000. Lecture 13: Graphs/Trees Cushion Treemaps. Jarke J. van Wijk and Huub van de Wetering, Proc nice abstraction work by authors themselves InfoVis 1999, pp 73-78. Online Dynamic Graph Drawing. Yaniv Frishman and Ayellet Tal. derived data: skeletonization via Strahler numbers Information Visualization Proc EuroVis 2007, p 75-82. Interactive Information Visualization of a Million Items. Jean-Daniel encoding techniques: CPSC 533C, Fall 2009 Fekete and Catherine Plaisant, Proc InfoVis 2002, p 117-124. ghosting = layering Topological Fisheye Views for Visualizing Large Graphs. Emden hiding = elision GrouseFlocks: Steerable Exploration of Graph Hierarchy Space. Daniel Gansner, Yehuda Koren and Stephen North, IEEE TVCG 11(4), p grouping = aggregation Archambault, Tamara Munzner, and David Auber. IEEE Trans. Tamara Munzner 457-468, 2005. Visualization and Computer Graphics 14(4):900-913 2008. UBC Computer Science Multiscale Visualization of Small World Networks. David Auber, Yves Chiricota, Fabien Jourdan, Guy Melancon, Proc. InfoVis 2003, p 75-81. Mon, 31 October 2011 Visual Exploration of Multivariate Graphs. Martin Wattenberg, Proc. [Fig 22. Herman, Melancon, and Marshall. Graph Visualisation in Information CHI 2006, p 811-819. Visualisation: a Survey. IEEE Transactions on Visualization and Computer Graphics, 6(1), pp. 24-44, 2000] 1 / 41 2 / 41 3 / 41 4 / 41 Trees: Basic Node-Link Drawings Trees: Basic Node-Link Drawings Trees: Radial Node-Link Drawings Trees: Radial Node-Link Drawings task/data abstraction algorithm level: animation requirements identified: data abstraction: data stream, not static file understanding detailed topological structure of tree Wetherell and Shannon 1978, Tidy Drawings of Trees encoding technique: radial not rectilinear layout avoid center collapse/clutter by interpolate polar not rectilinear visual encoding: layered node-link view Reingold and Tilford 1981, Tidier Drawing of Trees interaction technique: animated transitions from old to Walker 1990, A Node-positioning Algorithm for General vertical position: distance from root node in hops new layout Trees horizontal position: (as much symmetry as possible) Buchheim et al 2002, Improving Walker’s Algorithm to Run in Linear Time maintain neighbor order to stabilize (note prefuse bug!) [http://gravite.labri.fr/?Want to work with us ?:Hiring puzzles:Tidy Tree Layouts] [http://gravite.labri.fr/?Want to work with us [Figs 3, 5. Yee et al. Animated Exploration of Graphs with Radial Layout. Proc InfoVis 2001.] [Fig 2. Yee et al. Animated Exploration of Graphs with Radial Layout. Proc InfoVis 5 / 41 6 / 41 7 / 41 2001.] 8 / 41 Trees: Treemaps Cushion Treemaps Scaling Up Treemaps: MillionVis Scaling Up Treemaps: MillionVis visual encoding: treemaps, scatterplots interaction: animated transitions data abstraction: tree nodes have attributes visual encoding: also show nesting/topo structure more clearly with shading cues darkness shows nesting level visenc requirement: stable layout task abstraction: emphasize node attribs, not topological algorithm: many GPU tricks for speed structure interaction: scale parameter controls global vs. local dynamic queries through Z buffering visual encoding: use containment not connection [Fig 1. van Wijk and van de Wetering. Cushion Treemaps. Proc InfoVis 1999, pp 73-78.] [Figs 4, 5, 6. van Wijk and van de Wetering. Cushion Treemaps. Proc [Fig 4a. Fekete and Plaisant. Interactive Information Visualization of a Million Items. [Fig 1. Fekete and Plaisant. Interactive Information Visualization of a Million Items. [http://www.cs.umd.edu/hcil/treemap-history/treeviz colorful scaled.gif] InfoVis 1999, pp 73-78.] Proc InfoVis 2002, p 117-124.] Proc InfoVis 2002, p 117-124.] 9 / 41 10 / 41 11 / 41 12 / 41 Scaling Up Treemaps: MillionVis Graphs: Hierarchical Layout Graphs: Circular Layout Graphs: Force-Directed Placement scalability requires care at visual encoding level visual encoding visual encoding visual encoding not just algorithm level! vertical position: distance from root nodes on circle nondeterministic placement does not mean using containment algorithm to visually distinguish with fewer pixels, use shading not edge crossings minimized algorithms outline spring forces pull together edges, repulsive forces pull algorithms Sugiyama et al 1983, Methods for Visual Understanding apart nodes Six and Tollis 1999, A Framework for Circular Drawings of Hierarchical System Structures optimization framework easy to extend, but tends to be of Networks Gansner et al 1993, A Technique For Drawing Directed brittle Graphs algorithms Eiglsperger et al 2005, An efficient implementation of Fruchterman and Reingold, 1991, Graph Drawing By Sugiyama’s algorithm for layered graph drawing Force-Directed Placement Kamada and Kawai, 1989, An Algorithm For Drawing [Fig 2. Fekete and Plaisant. Interactive Information Visualization of a Million Items. General Undirected Graphs Proc InfoVis 2002, p 117-124.] 13 / 41 14 / 41 15 / 41 16 / 41

Online Dynamic Graph Drawing Online Dynamic GD: Algorithm Online Dynamic GD: Validation Critique data abstraction: streaming data not static file static graph layout algs unstable algorithm level small changes in input can have large changes in output complexity analysis task abstraction: dynamic stability (tradeoff) randomness, no constraints on maintaining geometric benchmarks: running time for CPU and GPU versions minimize visual changes proximity stay true to current dataset structure visual encoding level dynamic online algorithm qualitative discussion of result images/video first step: initialize, layout quantitative metrics: later steps: merge, pin, layout, animate pairwise avg node displacement for stability potential energy for quality acceleration: partition before GPU force-directed layout compare static, full dynamic, dynamic without pinning [Fig 1. Frishman and Tal. Online Dynamic Graph Drawing. Proc EuroVis 2007, p 75-82.] 17 / 41 18 / 41 19 / 41 20 / 41 Critique Critique Multi-level Graphs Multi-level Graphs: GrouseFlocks strengths strengths data abstraction: create cluster hierarchy on top of visual encoding: containment strong algorithmic contribution strong algorithmic contribution original graph (coarsening) interaction: expand/contract metanodes to change graph previous work not scalable previous work not scalable cut Graph Hier 1 Graph Hier 2 Graph Hier 3 very good validation, matches technique contribution very good validation, matches technique contribution best paper award, EuroVis 2007 best paper award, EuroVis 2007 weaknesses using mesh datasets to test graph drawing claims different topological characteristics than typical infovis case [Fig 3. Archambault et al. GrouseFlocks: Steerable Exploration of Graph Hierarchy [Fig 2. Archambault et al. GrouseFlocks: Steerable Exploration of Graph Hierarchy Space. IEEE Trans. Visualization and Computer Graphics 14(4):900-913 2008.] Space. IEEE Trans. Visualization and Computer Graphics 14(4):900-913 2008.] [Fig 3a. Frishman and Tal. Online Dynamic Graph Drawing. Proc EuroVis 2007, p 21 / 41 75-82.] 22 / 41 23 / 41 24 / 41 Small-World Networks Small World Coarsening Small World: Nested Quotient Graphs Small World: Nested Quotient Graphs high clustering, small path length visual encoding remove low-strength edges pro: very evocative of structure vs. random uniform distribution containment: subgraph laid out within metanode maximal disconnected subgraphs con: does not scale past 2-3 levels of depth examples quotient graph: subgraph = higher-level node social networks, movie actors, Web, ... multiscale small-world networks exploit these properties for better layout [Fig 2. Auber et al. Multiscale Visualization of Small World Networks. Proc. InfoVis 2003, p 75-81.] [Fig 5. Auber et al. Multiscale Visualization of Small World Networks. Proc. InfoVis 2003, p 75-81.] [Fig 3. Auber et al. Multiscale Visualization of Small World Networks. Proc. InfoVis 25 / 41 26 / 41 27 / 41 28 / 41 2003, p 75-81.] Topological Fisheye Views Topological Fisheye Views Topo Fisheye: Coarsening Strategy Topo Fisheye: Coarsening Requirements data abstraction must preserve graph-theoretic properties uniform cluster/metanode size input is laid-out graph topological distance (hops away), cycles match coarse and fine layout geometries construct multilevel hierarchy by coarsening graphs cannot just use geometric proximity alone scalable interaction: user controls focus point cannot just contract nodes/edges visual encoding: show hybrid view made from cut through exploit geometric information with proximity graph several levels [Fig 10. Gansner, Koren, and North, Topological Fisheye Views for Visualizing Large Graphs. IEEE TVCG 11(4), p 457-468, 2005.] [Fig 2. Gansner, Koren, and North, Topological Fisheye Views for Visualizing Large Graphs. IEEE TVCG 11(4), p 457-468, 2005.] [Fig 2. Gansner, Koren, and North, Topological Fisheye Views for Visualizing Large [Fig 4,7. Gansner, Koren, and North, Topological Fisheye Views for Visualizing Large Graphs. IEEE TVCG 11(4), p 457-468, 2005.] Graphs. IEEE TVCG 11(4), p 457-468, 2005.] 29 / 41 30 / 41 31 / 41 32 / 41