370 likes | 468 Views
Lecture 16 : Visualizations Summary. April 2, 2011 COMP 150-7 Visualization. Data Types. Dimensionality 1D , 2D 3D Temporal Multi-dimensional ( nD ) Relationships Tree (hierarchy) Network Others Workspaces Text. Data Types. Dimensionality 1D , 2D – boring
E N D
Lecture 16:Visualizations Summary April 2, 2011 COMP 150-7Visualization
Data Types • Dimensionality • 1D, 2D • 3D • Temporal • Multi-dimensional (nD) • Relationships • Tree (hierarchy) • Network • Others • Workspaces • Text
Data Types • Dimensionality • 1D, 2D – boring • 3D – take the graphics course • Temporal • Multi-dimensional (nD) • Relationships • Tree (hierarchy) • Space Filling (e.g., Tree maps) • Node-Link (e.g., org charts) • Network • Others • Workspaces – huh? • Text – important, but not covered. Why?
Visualization Examples • http://mbostock.github.com/protovis/ex/ • https://github.com/mbostock/d3/wiki/Gallery • http://www.tableausoftware.com/public/gallery • http://www-958.ibm.com/software/analytics/manyeyes/ • http://www.visual-literacy.org/periodic_table/periodic_table.html • TreeVis.net
Equivalence in Visualizations? • TreeVis.net examples • Sunburst -> Icicle Plot -> Circle Hierarchy • Circle Hierarchy -> (top-down) Tree • Circle Hierarchy -> Nested Bubbles -> Treemap • Sunburst -> Reingold–TilfordTree • Others? • LineGraph -> Barchart -> Stacked Barchart -> ThemeRiver • (Conceptually Wrong, but cool-looking: http://bl.ocks.org/mbostock/1256572) • (Note the Horizon Graph)
Challenges • Our short-term memory is terrible • Recall of specific event or instance from an animation is very difficult, especially when the visualization is abstract • Remember Hans Rosling’sGapminder • The questions are: • How to layout time? • Usually x-axis • Cultural differences? • How to find a pattern in time? • Visually? • Computation?
Challenges • Our brain cannot handle high dimensional information • We see the world in 2.5D, our brain reconstructs 3D • Dimensionality higher than that is too abstract to “visualize” • The questions are: • How to see more than 2 dimensions? • Slice-and-Dice • Stacking • How to find correlations in Data? • Visually? • Computation?
Challenges • Hierarchies occur everywhere, and can be used to represent lots of different things • A classic computer science issue • Problem is that hierarchies can be both “deep” and “wide” • The questions are: • How to utilize space effectively? • Space-filling • Node-link • How to find a pattern in the sub-branches? • Visually? • Computation?
Types of Hierarchy Visualizations • Space Filling • Node-Link
How to Draw This Tree? • How wide does this graph have to be?
2D Hyperbolic Browser http://flare.prefuse.org/demo
Other Space-Filling + Node-Link Graphs • Hilbert Curves
Challenges • Thinking about many-to-many relationships is always challenging • Students vs. classes • Particularly tricky because of: • Directed vs. undirected edges • Cycles within the graph • The questions are: • Hairball problem • Too many edges, too many nodes – layout is very challenging • Interaction is a must • How to find a pattern in a network? • Visually? • Computation?
What Makes a Graph? • Vertices (nodes) • Edges (links) • Adjacency list: • 1: 2 • 2: 1, 3 • 3: 2
Matrix Representation Henry InfoVis 06
NodeTrix Hybrid of matrix and node-link Best of both worlds? Worst of both methods? Henry InfoVis 07