Tree structured node-link diagrams grew too large to be useful, so Iexplored ways to show a tree in a space-constrained layout. I rejectedstrategies that left blank spaces or those that dealt with only fixed levels orfixed branching factors. Showing file size by area coding seemed appealing, butvarious rectangular, triangular, and circular strategies all had problems. Thenwhile puzzling about this in the faculty lounge, I had the Aha! experience ofsplitting the screen into rectangles in alternating horizontal and verticaldirections as you traverse down the levels. This recursive algorithm seemedattractive, but it took me a few days to convince myself that it would alwayswork and to write a six line algorithm. This algorithm and the initial designsled to the first Technical Report (HCIL TR91-03) in March 1991 which was published in the ACM Transactions onGraphics in January 1992 ( -11-1/p92-shneiderman/).Choosing the right name took probably as long, but the term 'treemap' describedthe notion of turning a tree into a planar space-filling map.
My initial design simply nested the rectangles, but a more comprehensibledesign used a border to show the nesting. Finding an effective visualizationstrategy took only a few months but producing a working piece of software tookover a year. Brian Johnson implemented the algorithms and refined thepresentation strategies while preserving rapid performance even with 5,000 nodehierarchies. The TreeViz applicationran on color Macintosh models and led to the widely cited paper (HCIL TR91-06) jointly authored paper in the October 1991 IEEE Conference onVisualization. This paper was reprinted in Readingsin Information Visualization. I think treemaps are a convenientrepresentation that has unmatched utility for certain tasks. The capacity tosee tens of thousands of nodes in a fixed space and find large areas orduplicate directories is very powerful. I still use TreeViz for cleaning up myMacintosh. It does take some learning for novices to grasp the tree structurelayout in treemaps, but the benefits are great.
My excitement about treemaps was great and like many innovators I thoughtmillions of users would be using this tool within a few years. Our minds werenot focused on getting a patent, since I thought this was more of a conceptthat a product. Brian's implementation of TreeViz was registered with theUniversity's Office of Technology Liaison which sought to distribute TreeViz.
We found that new users took 10-15 minutes to get acquainted with thetreemap display, so we began to explore improvements and training methods. Wewere impressed to examine thousands of nodes at 5-7 levels at once on thescreen, but novices did better seeing 20-50 nodes at 1-3 levels. We had tobring our training times up to about 15 minutes in order to demonstrate thestrong benefits of treemaps. Brian Johnson's dissertation (HCILTR 94-04) reports on two studies, that were never published, that sort outthe benefits of his treemaps. The 1992 HCIL VideoReportsand the 1993 HCIL VideoReports showed TreeViz in action. TreeViz for Apple Macintoshis available for free downloading.
A Japanese visitor, Asahi Toshiyuki, built his own innovative treemapinterface to implement the Analytical Hierarchy Process in decision making (HCILTR 94-08). Users could express their opinions of the relative merits of adecision choice (such as which site to chose for a factory) by pumping up areasfor their preferred choices, and pumping up the areas for importance of costs,availability of labor, tax breaks, etc. (Figure 5). The video demonstratesthese processes (HCILVideo Reports 1994, and HCILTR 95-04) and an empirical study showed users could succeed with this tool.
In my travels to lecture about our work, treemaps became a major topic.However, I ran into resistance when showing still images of our hard diskdirectories with thousands of nodes. Once at the University of Washingtonafter my talk produced a mixed reaction about treemaps, I asked my audience tofollow me down to one of their labs. I installed TreeViz and examined theirhard disk directories. I immediately spotted a problem, and with a few cluesthey could see for themselves that there were three copies of the same Ccompiler installed on this machine. The x-ray vision metaphor had proven to beeffective on this occasion. Similarly, at Apple Computers, my audience muchpreferred the dynamic queries demos of the HomeFinder, but I gave copies ofTreeViz to several interested attendees. The next day one of them reportedfinding many megabytes of useless information on their network servers.
Micro Logic Corp, a New Jerseycompany, sells a commercial product, called DiskMapper Windows machines based on the treemap idea. They have received great pressattention and awards for their product. The University of Marylandreceives a modest royalty on DiskMapper by way of a license agreement with theOffice of Technology Liaison.
Statisticians point out that the mosaic display, shown by Bertin, and othersis similar to the treemap concept. For fixed level hierarchies there is a greatsimilarity, but the gist of the treemap idea was intimately tied to thecomputerized implementation and user control panel for setting attributes. Adiscussion of mosaic displays and an implementation that allows rapid switchingof the hierarchy is in a Univ of Maryland graduate student project: CatTrees:Dynamic Visualization of Categorical Data Using Treemaps.
During 1997 an implementation was done in Delphiby Univ. of Maryland grad students Jerome Brown and ShaunGittens. It is called TreeMap97 ( ).This general purpose version of slice-and-dice treemaps was revised by ChrisNorth. You can download the software (updated version)and run the sample data set or import your own data sets.
During Summer 2000, the HCIL resumed work on treemaps, when RaghuveeraChalasani developed a Java version ( )that included dynamic queries sliders. By early 2001 we polished hisimplementation into Treemap 3.0 is licensed by the Univ.of Maryland's Office ofTechnology Liaison (screen shot). Treemap 3.0 includes dynamic queries to filter out unwanteditems, squarified layout (as well as slice-and-dice), improved input, infotips,and color/font controls. A demo version with five data sets is available forfree. A useful feature of Treemap 3.0 is that it also allows you to visualizethe contents of your Windows PC directory structure.
The cluster and squarified treemap algorithms are visually appealing in partbecause the rectangles are more square like. They avoid the thin rectangles inthe slice-and-dice algorithm, but sacrifice the alphabetic ordering of nodes.This creates an additional problem when leaf node sizes change because theposition of rectangles can alter dramatically. The goal of ordered treemaps wasachieved by a novel algorithm that nicely balanced square-like nodes (aspectratios close to one) while preserving order. This paper, "OrderedTreemap Layouts" was presented at the October 2001 IEEE Symposium onInformation Visualization.
However, we found still more ways to improve the layout by organizing thescreen space into horizontal (or vertical) strips. This idea helped keep thesquare-like aspect ratio and made for an easier to follow ordering. Wedemonstrated this with a study of 20 users. Another innovation, quantum treemaps,satisfy the need to accommodate fixed shape items like page thumbnails orphotos. The paper, "Orderedand Quantum Treemaps: Making Effective Use of 2D Space to Display Hierarchies"appeared in ACM Transactions on Graphics (October 2002, vol 21, no 4, pp833-854) ( =571647.571649). The figure below shows a quantum strip treemap of a photo image browser, PhotoMesa.
Treemap 4, released inFebruary 2003, includes a flexible hierarchy, color binning, improved colorsetting, aggregation, directory mapping, and export. Our applicationsinclude treemaps to show oil production data (SPEpaper), project management (PMIpaper), and a geneontology with 14,000 nodes at 16 levels. Our work with biologist Eric Baehrecke led to a specialversion of Treemap to handle the Gene Ontology and gene expression data from microarray studies(see paper in BMC Bioinformatics 2004,5:84)
In august 2003, substantial codemodifications were made to speed loading, permit use as an applet, andenhance usage of tm3 files. In addition, onlinedocumentation has been added in the usual textual format, and also as aset of 13 user selectable QuickDemos (animated with sound) that run from 11 to 75 seconds. A 9-minute videooverview is also available.
A 20-year review of treemaps as compared with hyperbolic and cone trees reveals patterns in researchcitations, trade press mentions, and patent data. Conjectures about why treemaps succeeded are made byShneiderman, B., Dunne, C., Sharma, P., and Wang, P., Innovation Trajectories forInformation Visualizations: Comparing Treemaps, Cone Trees, and Hyperbolic Trees, InformationVisualization Journal 11, 2 (2011), 87-105.The summer of 2013 was a good time for Ben Shneiderman to work with grad student Minhaz Kazi to producethe Treemap Art Project. The 12 large (24 x 36 inches)colorful images are framed and hanging in the3rd-floor hallway at the UM Computer Sciences InstructionalCenter. The designs and color palettes were influenced by modern art masters such as Mondrian, Albers,Davis, Hoffman, etc., but they still show real data with significant insights.Related Treemap Research and DevelopmentJohn Stasko at Georgia Tech produced a nice X-Windows version, and hiscolleagues Sougata Mukherjea, James D. Foley, and Scott Hudson produced a ACMCHI95 paper that used treemaps: Visualizing Complex Hypermedia Networks throughMultiple Hierarchical Views ( _bdy.htm)
Interesting treemap applications began sprouting during fall 2000. AtTruePeers youcan compare personalities with a cluster-style PeopleMap, and they include aspecial year 2000 election comparison among the 42 US Presidents. At Peets Coffeeyou'll find a neat treemap that shows their 32 coffees with area showing priceand color (from light to dark) showing flavor _selector.asp.
760c119bf3