Musings on Treemaps John Stasko Information Interfaces Research Group College of Computing / GVU Center Georgia Institute of Technology HCIL, Univ. of Md. May 31, 2001

2 Background zEarly UNIX implementations of Treemap XTango animation using technique Current UNIX tool

3 Treemap Variation zWhat if we used a radial expansion (with root at center) rather than a rectangular expansion? zAdd explicit representation of parent/directory

4 Sunburst Visualizing file and directory structures Root dir at center Color - file type Angle - file/dir size Quick demo

5 Appears in: American Heritage Dictionary, 3rd Ed. Houghton Mifflin, 1992

6 Empirical Study zCompare Treemap (borderless) and Sunburst with users performing typical file/directory-related tasks zEvaluate task performance on both correctness and time International Journal of Human-Computer Studies Special issue on Empirical Studies of InfoVis, 53(5), Nov. 2000, pp With R. Catrambone, M. Guzdial, K. McDonald

7 Experiment z60 participants zParticipant only works with a small or large hierarchy in a session zVary order across participants SB A, TM B TM A, SB B SB B, TM A TM B, SB A 32 on small hierarchies 28 on large hierarchies Small Hierarchy (~500 files) Large Hierarchy (~3000 files) A B A B

8 Tasks Identification (naming or pointing out) of a file based on size, specifically, the largest and second largest files (Questions 1-2) Identification of a directory based on size, specifically, the largest (Q3) Location (pointing out) of a file, given the entire path and name (Q4-7) Location of a file, given only the file name (Q8-9) Identification of the deepest subdirectory (Q10) Identification of a directory containing files of a particular type (Q11) Identification of a file based on type and size, specifically, the largest file of a particular type (Q12) Comparison of two files by size (Q13) Location of two duplicated directory structures (Q14) Comparison of two directories by size (Q15) Comparison of two directories by number of files contained (Q16) Representative of things people really do

9 Results (1) Small hierarchy A B Phase 1 TM 9.88 (3.23) TM (2.14) SB (1.96) SB (1.69) Phase 2 TM (1.75) TM (2.77) SB (2.00) SB (2.00) Correct task completions (max=16)

10 Results (2) Large hierarchy A B Phase 1 TM 8.71 (1.60) TM 8.29 (2.14) SB (1.27) SB (2.67) Phase 2 TM (1.27) TM (1.57) SB (2.16) SB (2.00) Correct task completions (max=16)

11 Results (3) zOrdering/learning effect for Treemap on large hierarchies zPerformance trends favored Sunburst, but not clear-cut (completions, time) zSubjective preference: SB (51), TM (9), unsure (1)

12 Observations zSB appeared to convey structure better zParticipants felt TM conveyed size better, but not bore out by data zStrategies mattered

13 SunBurst Negative zIn large hierarchies, files at the periphery are usually tiny and very difficult to distinguish examples

14 Treemap Too examples

15 Fix: Objectives zMake small slices bigger zMaintain full circular space-filling idea zAllow detailed examination of small files within context of entire hierarchy zDon’t alter ratios of sizes zAvoid use of multiple windows or lots of scrollbars zProvide an aesthetically pleasing interface in which it is easy to track changes in focus

16 3 Solutions zThree visualization+navigation techniques developed to help remedy the shortcoming yAngular detail yDetail outside yDetail inside Proceedings of Information Visualization 2000 Oct. 2000, pp With E.Zhang

17 Design 1 - Angular Detail Most “natural” Least space-efficient Most configurable by user

18 Design 2 - Detail Outside Exhibits non-distorted miniature of overview Somewhat visually disconcerting Focus is quite enlarged (large circumference and 360°) Relatively space efficient

19 Design 3 - Detail Inside Perhaps least intuitive and most distorting Items in overview are more distinct (larger circumference) Interior 360° for focus is often sufficient

20 Video 4 minutes

21 Key Components zTwo ways to increase area for focus region: larger sweep angle and longer circumference zSmooth transitions between overview and focus allow viewer to track changes zAlways display overview zAllow focus selections from anywhere: normal display, focus or overview regions

22 Implementation zUtilizes fundamental animation update routine yExample: Detail Outside (called 3 times) xShrink global view xFocus region grows out xFocus regions wraps around global view ySmooth interpolation between start-end position and angle

23 Speed Considerations zDon’t draw small slices zCache small and large images of entire hierarchy, reload rather than draw zDuring animation transitions, only draw the 100 largest slices (don’t use thresh- holding) -> Consistent speed as hierarchy grows (really dependent on processor & graphics)

24 Treemap Application zCan similar techniques be applied to treemaps to support focusing on small items? yExpand area given to focus item ySmoothly animate focus change yChallenge: Where to position focus and context?

25 Idea 1: Superimpose +: Simple - : Obscures areas

26 Idea 2: Grow/Shrink +: Shows all context - : Empty areas

27 Idea 3: Shrink overview +: Shows all context - : Wasteful of space

28 Thoughts? zOther possibilities? zWhich do you like best? zSomebody build one!

29 Summary zTreemaps can be a very helpful information visualization zTechniques for examining small regions in more detail can make them even more powerful

30 For More Information... zwww.cc.gatech.edu/gvu/ii