SIMS 247 Lecture 7 Simultaneous Multiple Views February 10, 1998 Marti Hearst SIMS 247
Dynamic Queries Instead of a formal database language Explore a dataset interactively Use graphical devices to interactively update a visualization Marti Hearst SIMS 247
Revisit VISAGE Roth et al.: apply visualization operations to data and metadata directly convert tabular to graph to map and back Pros: flexible much work done for user Cons: layout may be confusing which operation applied when? How to better organize views of results of operations? Marti Hearst SIMS 247
Tufte: Multiples in Space and Time Multiple images reveal: repetition and change pattern and surprise Create visual lists of objects & activities Help viewers compare analyze differentiate Narrate sequences of motion Marti Hearst SIMS 247
Multiple Views: Star Plot (Discussed in Feinberg 79 Multiple Views: Star Plot (Discussed in Feinberg 79. Works better with animation. Example taken from Behrans & Yu 95.) Marti Hearst SIMS 247
One Approach to Many Simultaneous Views Anupam et al. 97: Show many views on database simultaneously Example: telecomm service Rows: services Columns: companies Use transparency to fully or partially overlay different graphs (see pictures in class) Marti Hearst SIMS 247
Another Approach: Linked Spreadsheet Cells (Chi 97) Spreadsheet: familiar interface Items are compared side-by-side Operations easily applied across groups operations on rows or columns thus differs from scatterplot matrix The twist: cells contain results of a large amount of data Different operators for different datasets Marti Hearst SIMS 247
Spreadsheet Algorithm Example (Chi 97) Problem: Delauny triangulation generate 3D points randomly try to find volumes the algorithm is complex Columns: how many steps the algorithm has run (5, 6, 25, 50) Rows: 1) points (scatterplot) 2) show planes transparently 3) show edges between points 4) result of several operations Marti Hearst SIMS 247
Visualizaiton Spreadsheet; Chi 97 Marti Hearst SIMS 247
Molecular Biology Spreadsheet Problems with matrices: identify differences between matrices decide which visualization to apply Goal: how do the datasets differ? rows: different data sets columns: 1) positive values -> cube points 2) all values -> carpet (red = negative) 3) all values -> bars 4)different rotational views Claim: easy to explore application of various operators Marti Hearst SIMS 247
Visualizaiton Spreadsheet; Chi 97 Marti Hearst SIMS 247
How to Specify Operators on Data? Selecting from menu in the spreadsheet is awkward Suggestion: Tool Glass/Magic Lenses (Stone, Fishkin, Bier, others) drag a translucent operator palette around combine operators directly on the palette two-handed interface How might they be applied to linking in visualization? Marti Hearst SIMS 247