Full text of DOI:10.1109/38.814540

Vision 2000
Op p ort unit ies for
Inform at ion
graphical properties of points, lines, and areas. For
example, Figure 2 shows a tree map, a visualization
invented by Ben Shneiderman for hierarchical data that
subdivides a quantitative value.5 In this case, the area
of rectangles represents the capitalization of companies
on the stock market.6 Green (red) shows positive (neg-
Visualizat ion
nformation visualization has been an active ative) changes in price.
I
research topic for more than a decade.1
Products are starting to appear. This article discusses Visualizat ion an d com p ut at ion
some of the research and commercial opportunities for
Computation creates content. Visualization has an
information visualization over the next decade.
opportunity to make computation and its content acces-
sible to humans. For example, visual data mining uses
visualization to augment data mining, which uses algo-
Visualizat ion an d p resen t at ion
Visualization is related to presentation, which has its rithms to find patterns in databases.7 These algorithms
roots in the work of William Playfair, an English politi- are difficult for decision makers to use. Visualization can
cal economist who in 1786 published the first known make the data and the patterns more accessible, allow-
time-series.2 Presentation is primarily about using visu- ing comparison and verification of results.8 Visualization
al representations to communicate. In 1967, the French can also be used to steer the data mining algorithms.9
cartographer Jacques Bertin emphasized that visual rep-
resentations should also be used to think.3 Visualiza- Hum an p ercep t ion
tion, in particular, is about using computer graphics to
The power of visualization to exploit human percep-
think about more cases, more variables, more relations. tion offers both a challenge and an opportunity. The
The goal is to think clearly, appropriately, with insight, challenge is that incorrect patterns can be perceived in
and with conviction to act.
visualizations, leading to incorrect decisions and
Unlike presentations, visualizations are typically inter- actions. The opportunity is to use knowledge about
active and animated. For example, Figure 1 shows a screen human perception when designing visualizations. How-
shot of the cone tree, a 3D visualization of hierarchical ever, existing knowledge about human perception and
data.4 Hierarchies have the problem that they generally presentation design is insufficient. Visualization creates
grow wider faster than they grow deep, creating an awk- a feedback loop between perceptual stimuli and the
ward aspect ratio. The cone tree tames this ratio with 3D user’s cognition. A hot topic for research in the coming
graphics and interactive animation. Although many nodes decade will be characterizing this feedback loop, under-
are occluded in the static image in Figure 1, interactive standing the cognitive impact of visualization and the
animation provides rapid access to those nodes and fos- perceptual impact of cognition.
ters a percept of them even when occluded.
Jock D.
Mackinlay
Xerox Palo Alto
Research
Center
3D g rap h ics
3D graphics offers a controversial opportunity for
In form at ion visualizat ion
Visualization divides roughly into two areas, depend- information visualization, even though it’s often used
ing on whether physical data is involved. Scientific visu- for physical data in scientific visualization. Nonphysical
alization, discussed elsewhere in this issue, focuses data, such as the hierarchical data in Figures 1 and 2,
primarily on physical data such as the human body, the can be shown in either 2D or 3D. Although 3D is visual-
earth, molecules, and so on. Information visualization ly exciting and provides an additional dimension for
focuses on abstract, nonphysical data such as text, hier- showing data, it often requires specialized graphics
archies, and statistical data.
hardware. Furthermore, the perspective and texture
The challenge for nonphysical data is designing a visu- mappings degrade font quality, which is often a key
al representation. One approach is to map the non- aspect of information visualization. However, the next
physical data to a virtual object such as a cone tree, decade will bring improved software graphics perfor-
which can be manipulated as if it were a physical object. mance and high-resolution LCD displays that will make
Another approach is to map the nonphysical data to the 3D more practical for information visualization.
22
January/ February 2000
0272-1716/00/$10.00 © 2000 IEEE

Visualizat ion for e-com m erce
Visualization has the potential to help e-commerce
customers manage the overwhelming choices available
on the Web. Consider eBay, which has thousands of
pages describing auctions and commenting on the reli-
ability of participants. The effort required to click and
scroll through these pages makes comparison shopping
difficult. Another example is any e-commerce site that
provides a one-stop shopping service. Their success
depends partially on giving the customer an unbiased
impression of their product offerings. Visualization can
be used in both of these examples to give potential cus-
tomers a powerful tool for comparison shopping. Rapid
interaction can lead to a confident consumer and hence
a sale.
1 Cone tree of
10,000 pages
from the PARC
Web site.
In t eg rat ion
Over the past decade, research on information visual-
ization has focused on developing specific visualization
techniques.10 An essential task for the next decade is to
integrate these techniques into larger systems that sup-
port information work, which has three steps: (1) forag-
ing for data, (2) thinking about data, and (3) acting on
data. With attention to standards for connecting visual-
izations to databases, computation, and other visualiza-
tions, visualization can support all aspects of this process.
Foraging can be supported by making databases visible.
Thinking about data can be supported by choosing the
appropriate visualization, which can be done in a spe-
cialized visualization called an information workspace.10
Finally, actions can be supported by integrating visual-
izations into the user interfaces of applications.
2 Market map,
a tree map of
market value.
Con clusion
3. J. Bertin, Semiology of Graphics: Diagrams, Networks, Maps,
University of Wisconsin Press, Madison, Wis., 1967, repub-
lished 1983.
4. G.G. Robertson, J.D. Mackinlay, and S.K. Card, “Cone
Trees: Animated 3D Visualizations of Hierarchical Infor-
mation,” Proc. ACM Conf. on Human Factors in Computing
Systems (CHI 91), ACM Press, New York, 1991, pp.
189–194.
5. B. Johnson and B. Shneiderman, “Tree Maps: A Space-Fill-
ing Approach to the Visualization of Hierarchical Infor-
mation Structures,” Proc. IEEE Visualization 91, IEEE
Computer Society Press, Calif., 1991, pp. 284–291.
6. M. Wattenberg, “Visualizing the Stock Market,” Conf. Com-
panion of CHI 99, ACM Press, New York, 1999, pp. 188-189.
7. P.C. Wong, “Visual Data Mining,” IEEE CG&A, Vol. 19, No.
5, Sep./Oct. 1999, pp. 20-21.
The prospect for information visualization in the mar-
ketplace looks promising. However, external issues
often make it difficult to time the movement of research
to products. In particular, the Web has created a client-
server model for disseminating information that works
poorly for information visualization, which requires
high-bandwidth access to data. Unfortunately, the mar-
ket encourages user interface design to focus on the low-
est common denominator, which currently doesn’t
include technology that could deliver information visu-
alization products. Although these obstacles make the
precise timing difficult to predict, advances in hardware
over the coming decade will surely move information
visualization from research to products.
■
Acknowledgments
8. A. Pang and H.-G. Pagendarm, “Visualization for Every-
one,” IEEE CG&A, Vol. 18, No. 5, Jul./Aug. 1998, pp.
47-48.
9. A. Hinneburg, D.A. Keim, and M. Wawryniuk, “HD-Eye:
Visual Mining of High-Dimensional Data,” IEEE CG&A, Vol.
19, No. 5, Sep./Oct. 1999, pp. 22-31.
I thank the following people for their helpful com-
ments and suggestions: Polle Zellweger, Allison
Woodruff, Ed Chi, Stuart Card, Ben Shneiderman,
George Robertson, and Tamara Munzner.
10. S.K. Card and J.D. Mackinlay, “The Structure of the Infor-
mation Visualization Design Space,” Proc. IEEE Symposium
on Information Visualization (InfoVis 97), IEEE Computer
Society Press, Los Alamitos, Calif., 1997, pp. 92-99.
References
1. S.K. Card, J.D. Mackinlay, and B. Shneiderman, Readings
in Information Visualization: Using Vision to Think, Morgan
Kaufman, San Francisco, 1999.
2. W. Playfair, The Commercial and Political Atlas, London, Contact Mackinlay at mackinla@parc.xerox.com.
1786.
IEEE Computer Graphics and Applications
23