What is Visualization Really for?


MIN CHEN, University of Oxford
LUCIANO FLORIDI, University of Hertfordshire and University of Oxford
RITA BORGO, Swansea University

Whenever a visualization researcher is asked about the purpose of visualization, the phrase “gaining
insight” by and large pops out instinctively. However, it is not absolutely factual that all uses of
visualization are for gaining a deep understanding, unless the term insight is broadened to encompass all
types of thought. Even when insight is the focus of a visualization task, it is rather difficult to know what
insight is gained, how much, or how accurate. In this paper, we propose that “saving time” in
accomplishing a user’s task is the most fundamental objective. By giving emphasis to saving time, we can
establish a concrete metric, alleviate unnecessary contention caused by different interpretations of insight,
and stimulate new research efforts in some aspects of visualization, such as empirical studies, design
optimisation and theories of visualization.
General Terms: Visualization

1. INTRODUCTION
Visualization was already an overloaded term, long before it has become a
fashionable word in this era of data deluge. It may be used in the context of
meditation as a means for creative imagination, or in sports as a means for creating a
heightened sense of confidence. If we consider the term literally, as Robert Spence
said, “visualization is solely a human cognitive activity and has nothing to do with
computers” [Spence 2007].
In this article, we focused on visualization in computing, which may be referred to
technically as Computer-supported Data Visualization. In this context, the process of
visualization features both data and computer. These two essential components
differentiate this technological topic from those above-mentioned contexts. In the
remainder of this article, we will simply refer to “computer-supported data
visualization” as “visualization”.
Scott Owen [1999] compiled a collection of definitions and rationale for
visualization, most of which are still widely adopted or adapted today. These
definitions were intended to define the two questions, namely what is visualization
and what is it for?
 “The goal of visualization in computing is to gain insight by using our visual
machinery.” [McCormick et al. 1987]
 “Visualization is a method of computing. It transforms the symbolic into the
geometric, ... Visualization offers a method for seeing the unseen. It enriches
the process of scientific discovery and fosters profound and unexpected
insights.” [McCormick et al. 1987]
 “Visualization is essentially a mapping process from computer representations
to perceptual representations, choosing encoding techniques to maximize
human understanding and communication.” [Owen 1999]
 “Visualization is concerned with exploring data and information in such a way
as to gain understanding and insight into the data. The goal ... is to promote a


Authors’ addresses: M. Chen, Oxford e-Research Centre, University of Oxford, UK; R. Borgo, Department
of Computer Science, Swansea University, UK; L. Floridi, Department of Philosophy, University of
Hertfordshire, UK.

The first draft of this article was completed on 20 February 2013

Page 2 Min Chen, Luciano Floridi, and Rita Borgo

deeper level of understanding of the data under investigation and to foster new
insight into the underlying processes, relying on the humans’ powerful ability
to visualize”, [Earnshaw and Wiseman 1992]
 “The primary objective in data visualization is to gain insight into an
information space by mapping data onto graphical primitives.” [Senay and
Ignatius 1990]
In addition to Scott Owen’s collection, there are other commonly cited definitions:
 Visualization facilitates “the use of computer-supported, interactive, visual
representations of abstract data to amplify cognition.” [Card et al. 1999]
 “Graphics reveal data. Indeed graphics can be more precise and revealing than
conventional statistical computations.” [Tufte 2001]
 “Information visualization helps think.” [Few 2009]
 “Information visualization utilizes computer graphics and interaction to assist
humans in solving problems.” [Purchase et al. 2008]
 “The goal of information visualization is to translate abstract information into
a visual form that provides new insight about that information. Visualization
has been shown to be successful at providing insight about data for a wide
range of tasks.” [Hearst 2009]
 “The goal of information visualization is the unveiling of the underlying
structure of large or abstract data sets using visual representations that utilize
the powerful processing capabilities of the human visual perceptual system.”
[Berkeley 2010]
 “The purpose of visualization is to get insight, by means of interactive graphics,
into various aspects related to some processes we are interested in ...” [Telea
2008]
In the above definitions, there are many references to gaining insight, or likewise
phrases such as amplifying cognition, seeing the unseen, unveiling structure,
answering questions, solving problems, and so forth. It is unquestionable that these
are the benefits that visualization can bring about in many occasions. There has been
an abundance of evidence to confirm such goals are achievable. However, insight is a
non-trivial concept. It implies “accurate and deep intuitive understanding” according
to many dictionaries. While it is what everyone who creates or uses visualization is
inspired to achieve, it is an elusive notion and rather difficult to measure, evaluate,
or validate objectively.
Perhaps it is also because of its vagueness, it is relatively easier for people to
interpret the term insight differently. The charged debate about chart-junks a few
years ago was perhaps partly caused by the diverse interpretation of what insight to
be gained from visualization.
The debate started with a paper by Bateman et al. [2010], which reported an
empirical study on the effects of using visual embellishments in visualization. They
compared conventional plain charts with highly embellished charts drawn by Holmes
[1984]. The findings of the study suggest that embellishment may aid memorization.
Following this work, Hullman et al. [2011] proposed a possible explanation that
“introducing cognitive difficulties to visualization” “can improve a userʼs
understanding of important information.” Obviously this was a major departure from
the traditional wisdom of avoiding chart-junks in visualization. For example, in
[Tufte 2001], some of Holmes’s visual designs were shown as counter examples of this
wisdom.

The first draft of this article was completed on 20 February 2013

What is Visualization Really for? Page 3

These two pieces of work attracted much discussion in the blogosphere. Stephen
Few, the author of several popular books on visualization (e.g., [Few 2009]), wrote
two articles. On [Bateman et al. 2010], he concluded:
“At best we can treat the findings as suggestive of what might be true, but not
conclusive.” [Few 2011a]
Few was much more critical on [Hullman et al. 2011]:
“If they’re wrong, however, which indeed they are, their claim could do great
harm.” [Few 2011b]
In many ways, the two sides of the debate were considering different types of
insight to be gained in different modes of visualization. We will revisit this debate
later in Section 3.2.

2. A STORY OF LINE GRAPH

Before we attempt to answer the question what visualization is really for, let us
examine some examples of visualization. We start with one of the simplest form of
visualization, line graph, which is also referred to as line chart and line plot.

Fig. 1. This is the earliest line graph found in the literature. It divides the 2D plane onto some
30 temporal zones across the x‐axis and uses horizontal lines to indicate zodiac zones across
the y‐axis. Seven time series were displayed in this chart. Source: [Funkhouser 1936].

Fig. 1 shows a line graph created by an unknown astronomer in the 10th (or
possibly 11th) century, depicting the “inclinations of the planetary obits as a function
of the time” [Funkhouser 1936]. More line graphs were found in the 17th century
records, noticeably the plot of “life expectancy vs. age” by Christiaan Huygens in
1669, and the plot of “barometric pressure vs. altitude” by Edmund Halley in 1686
[Friendly2007]. The 18th and 19th centuries saw the establishment of statistical
graphics as a collection of charting methods, attributed to William Playfair, Francis
Galton, Karl Pearson and others [Cleveland 1985]. The invention of coordinate
papers in the 18th century also helped make line graph a ubiquitous technique in
science and engineering.
Today, digitally stored data that captures or exhibits a functional relationship y =
f(x) is everywhere. For example, there are thousands or millions of real time data
feeds of financial information. Weather stations and seismic monitors around the
world generate an overwhelming amount of data in the form of y = f(t). In some cases,

The first draft of this article was completed on 20 February 2013

Page 4 Min Chen, Luciano Floridi, and Rita Borgo

we still use line graphs for visualization, and in other cases, we do not. What has
been the most fundamental factor that makes visualization users choose one visual
representation from another? Is it a more quantifiable factor, such as the number of
data series, the number of data points per data series, or another data-centric
attribute? Is it a less quantifiable factor such as the amount or type of insight, the
amount of cognitive load required, the level of aesthetic attraction, the type of
judgment to be made, or any other human-centric attribute?
Let us consider why a seismologist uses a seismograph, which is a type of line
graph that depicts the measured vibrations over time. (For the convenience of
referring, we use female pronouns for the seismologist.) The main task supported by
a seismograph is for a seismologist to make observation. Her first priority is simply to
see, or to know, the data stream in front of her, so she can confidentially say “I have
seen the data”. She may wish to observe some signature patterns of a potential
earthquake, relationships among several data series measured at different locations,
anomalies that may indicate malfunction of a device, and so on. The seismologist also
uses seismographs as a mechanism of external memorization, since they “remember”
the data for her. In real time monitoring, she does not have to stare at the
seismometer constantly and can have a break from time to time. In offline analysis,
she does not need to remember all historical patterns, and can recall her memory by
inspecting the relevant seismographs. Viewing seismographs simulates various
thoughts, such as hypotheses. After observing a certain signature pattern in a
seismograph, she may hypothesise that the vibrations would become stronger in the
next few hours. While the seismograph advances with newly arrived data, she
evaluates her hypothesis intuitively. When discussing with her colleagues, she draws
their attention to the visual patterns on the seismograph, and explains her
hypothesis and conclusion. In other words, she uses the seismograph to aid her
communication with others.
Perhaps the seismologist does not have to use seismographs. The vibration
measures could simply be displayed as a stream of numbers; after all viewing these
numbers would be more accurate than viewing the wiggly line on a seismograph.
Alternatively, to make more cost-effective use of the visual media, the stream of
number could be animated in real time as a dot moving up and down, accompanied
by a precise numerical reading updated dynamically. Let us have a close look at the
advantages of a seismograph over a stream of numbers or an animation.

2.1 Making Observation

It is not difficult for most people to conclude that viewing a stream of numbers is
much slower than viewing a line graph such as a seismograph. Numerous studies in
psychology have confirmed this (e.g., [Styles 2006]). The latter often facilitates pre-
attentive processing, allowing information to be obtained from a visual medium or
environment unconsciously. One might suggest that line graphs make better use of
space than a stream data. This is certainly true, but space optimisation cannot be the
fundamental factor, as line graphs are not the optimal in terms of space usage in
static visualization [Chen and Jänicke 2010]. Furthermore, the animation of dots and
numbers would offer much better space utilisation.
The difficulty of using animation to support tasks of making observations is due to
its excessive demand for various cognitive capabilities, including attention and
memory. While watching such an animation, it is difficult for a viewer to pay
attention to a specific temporal pattern, and almost impossible to have a
photographic memory to record a specific set of numbers. Of course, one could view

The first draft of this article was completed on 20 February 2013

What is Visualization Really for? Page 5

the same animation repeatedly, and would eventually work out interesting patterns
in the movement of the dot and the variations of the numbers. It is no doubt much
slower than viewing a line graph.

2.2 Facilitating External Memorisation

A stream of numbers and an animation can facilitate external memorisation. In fact,
almost all digitally-stored data can do so. Hence the question should focus on how
fast a form of visual display can facilitate memory recall. Similar to what discussed
in Section 2.1, viewing a line graph is much quicker than viewing a stream of
numbers, or an animation. For example, if a seismologist tries to recollect her
memory about some events taking place over the past few hours, it only takes a few
seconds for her to trace her eyes along the seismograph to be reminded about what
happened before. It would perhaps take hours to read through thousands of numbers,
or to watch the animation repeatedly.

2.3 Stimulating Hypotheses and Other Thoughts

While some visualization tasks may involve routine, and perhaps sometimes
mundane, observations, others can be highly analytical, involving various aspects of
thought process, such as data comprehension, facts deliberation, experience
reflection, hypothesis generation, hypothesis evaluation, opinion formulation, and
decision making. Many aspects of human thinking are stimulated by visual signals.
If we compare a seismograph with a stream of numbers or an animation of dots and
numbers, we are interested in which type of visual signal can stimulate more
thought, or stimulate a specific aspect of thought faster. To our knowledge, there is
yet any reported study on such questions. However, with the availability of
technologies, such as functional magnetic resonance imaging (fMRI), we hope that
there will be more conclusive answers in the near future.
Nevertheless, there have been many anecdote evidences suggesting that when
visualization is appropriately designed to convey overviews and is supported by
interaction for details-on-demand exploration, it can simulate hypotheses more
effectively.

2.4 Evaluating Hypothesis

There is no doubt that hypothesis testing is a critical process in scientific
investigation. Whenever applicable and feasible, one should always utilise scientific
methods, such as statistical hypothesis testing and Bayesian hypothesis testing.
However, such scientific methods often require a non-trivial amount of time and
effort for collecting, processing and analysing data. In practice, visualization is often
used as an intuitive form of hypothesis evaluation. For example, in scientific
computation, to evaluate a simulation model (which is essentially a hypothesis),
scientists visualize the results of simulation and visually compare the results with
some ground truth data. Such intuitive evaluation is based on the principle
counterfactual reasoning. It is not in any way unscientific. It saves time. In the
visualization literature, there have been many case studies that confirm the use of
visualization as a tool for hypothesis evaluation.

2.5 Disseminating Knowledge

Visualization is used extensively for disseminating knowledge. In fact, this is often
mistaken as the main or only function of visualization. In such situations,
visualization is a tool for assisting a scientist or scholar in delivering a collection of

The first draft of this article was completed on 20 February 2013

Page 6 Min Chen, Luciano Floridi, and Rita Borgo

messages to an audience. These messages may consists of data being visualized,

background information to be appreciated, concepts to be comprehended, and
opinions to be accepted. Clearly, the person presenting the visualization would like to
direct the audience to receive the intended messages as fully, and as fast, as possible.
There is a subtle difference between this visualization task and those in Sections
2.1-2.4. Assessing how well a task is performed is generally difficult in many
practical situations. For example, consider a task of making seismological
observation. If a visual pattern of a potential risk was not noticed in a seismograph
during routine monitoring, unless the risk is actualised, it always seems debatable as
to such pattern should be noticed or not. The same paradox can be suggested for
external memorisation, hypothesis simulation and hypothesis evaluation. In
knowledge dissemination, however, as the person presenting the visualization
usually has a set of defined criteria for measuring task performance, he/she can
assess the audience to determine whether the intended messages were received.
Meanwhile, the time is more a constraint rather than a quality metric, since, for
instance, a presentation, a meeting or a lecture is usually time-limited. In such a
situation, visualization is often embellished in order to “energise” the messages
intended by the presenter.

3. SAVING TIME IN DIFFERENT MODES OF VISUALIZATION

The story of line graph in Section 2 highlights the importance of saving time in
performing a user’s tasks. Of course, this is not a new discovery. Amid many “insight-
based” definitions, some appreciated the purpose of saving time:
 “Today’s researchers must consume ever higher volumes of numbers ... If
researchers try to read the data, ... they will take in the information at snail’s
pace. If the information is rendered graphically, however, they can assimilate
it at a much faster rate.” [Friedhoff and Kiely 1990]
 “One of the greatest benefits of data visualization is the sheer quantity of
information that can be rapidly interpreted if it is presented well.” [Ware 2004]
 “Visual representations and interaction technologies provide the mechanism
for allowing the user to see and understand large volumes of information at
once.” [Thomas and Cook 2005]
 “Information visualization promises to help us speed our understanding and
action in a world of increasing information volumes.” [Card, 2007]

3.1 What is Visualization Really for?

DEFINITION. Visualization (or more precisely, computer-supported data visualization)
is a study of transformation from data to visual representations in order to facilitate
effective and efficient cognitive processes in performing tasks involving data. The
fundamental measure for effectiveness is correctness and that for efficiency is the
time required for accomplishing a task.
Note that we choose the verb “accomplish” to emphasise that the task has to be
performed to a certain degree of satisfaction before the measure of efficiency becomes
meaningful. When the correctness has reached a satisfactory level, or becomes
paradoxically difficult to assess (as discussed in Section 2.5), the time required to
perform a visualization task becomes the most fundamental factor. Such time is a
function of three groups of variables:
(a) data centric attributes, such as the size of a dataset, the number of
multivariate dimensions, the entropy of the data space, etc.
The first draft of this article was completed on 20 February 2013
What is Visualization Really for? Page 7

(b) human-centric attributes, such as amount or type of insight to be gained, the

type of judgment to be made, the amount of cognitive load required, the level of
aesthetic attraction, etc.
(c) information delivery attributes, such as the type of medium, the properties of
the display device, the type of visual representations, the type of exploration,
etc.
In most real world applications, there is usually little flexibility with (a) and (b).
Hence choosing the appropriate information delivery attributes can be critical to
accomplish a task efficiently.

3.2 Modes of Visualization

Visualization serves as a medium and a tool for human-human interaction. Let us
refer to those who create visualization as visualization producer and those who
view visualization in order to gain an insight as visualization consumer.
In some cases, the producer differs from the consumer. For example, a business
analyst, who has a good understanding of a financial data set, creates a collection of
charts for a company board meeting; or a teacher, who has a good understanding of a
concept, creates an illustration to disseminate his or her knowledge to students. In
many cases, the producer is also the consumer. For example, in a visual data mining
process, an analyst, who has difficulty to comprehend a complex data set by simply
reading the textual or numerical data, interactively explores various visual
representations of the data, in order to gain an overview or make a discovery.
Let us consider three types of visualization users: analyst A, who is a producer as
well as a consumer, presenter P, who is a producer but not a consumer, and viewer
V, who is a consumer but not a producer. Different combinations of analysts,
presenter and viewers in visualization processes will usually lead to different styles
of human-human interaction. Table 1 lists several typical operational modes of
visualization processes.
Table 1: Examples of common modes of human participation in visualization processes.
Mode Participants Example Scenarios
(1) A an analyst works alone.
(2) A1, A2, ..., Ak a team of analysts conduct collaborative visual data mining.
(3) A, V a personal visualization assistant and a boss.
(4) P, V a personal tutor and a student.
(5) P, V1, V2, ..., Vn a presenter (or a teacher) and an audience (or students).
(6) A1, A2, ..., Ak, a team of analysts carry out visual data mining in real time,
V1, V2, ..., Vn while a panel of onlookers eagerly observe the process.
(7) A, P, V1, V2, ..., Vn an analyst working for a domain expert who needs to
disseminate his/her research to others

Most analytical tasks (Sections 2.1-2.4) are likely to be conducted in modes (1), (2),
and (3). Only the tasks of knowledge dissemination are normally conducted in modes
(4) and (5). Mode (6) is relatively rare, but one can easily imagine that some
visualization tasks during disaster management may be performed in this mode. On
the other hand, mode (7) is rather common, but often has conflicting requirements
between the knowledge dissemination task and those analytical tasks.

The first draft of this article was completed on 20 February 2013

Page 8 Min Chen, Luciano Floridi, and Rita Borgo

3.3 Reasoning about Visual Embellishment

Let us revisit the debate about visual embellishment discussed in Section 1. Borgo et
al. [2012] reported a study on visual embellishment, which used more conservative
stimuli than [Bateman 2010]. It shows that visual embellishment may help
information retention in terms of both accuracy of and time required for memory
recall. However, this is at the expenses of an increase in the time required for visual
search, which is an activity typically taking place in tasks of making observation,
hypothesis generation and evaluation. Their study also indicates that visual
embellishment may help viewers to grasp concepts that a presenter would like to
disseminate, especially when such concepts are embedded among complex
information.
Borgo et al [2012] pointed out in their explanation that the finding about the
negative impact on visual search tasks provides scientific evidence to indicate some
disadvantages of using visual embellishments. In other words, visual embellishment
is unlikely to save time for an analyst in performing tasks such as making
observation, hypothesis generation and evaluation. They also pointed out that the
positive impact on memory and concept grasping should not be generalized to
situations where visualizations are created by data analysts for their own use. In
other words, the positive impact is relevant mainly to the above-mentioned modes (4)
and (5). In addition, they made a connection between their findings and the
information theoretic framework of visualization [Chen and Jänicke 2010].
Consider those visualization tasks discussed in Section 2. There are analytical
tasks (Sections 2.1-2.4), and dissemination tasks (Section 2.5). Table 2 summarises
the main characteristics of these two groups of visualization tasks. Since the majority
of work in visualization concerns about analytical tasks, the notion of “saving time”
must not sit on the backbench in the definition of visualization. As it implicitly
implies the completion of task, it encapsulates the notion of “gaining insight” to a
large degree, but not vice versa. Hence, “saving time” is more fundamental.
Table 2: Characteristics of analytical tasks and dissemination tasks in visualization.
Analytical Tasks Dissemination Tasks
Modes of visualization producer (may) = consumer producer  consumer
Saving time producer & consumer’s time producer & consumer’s time
Gaining insight for producer to gain for consumer to gain
Assessing correctness relatively difficult more feasible
Using embellishment usually not helpful can be helpful
Information theory source encoding channel encoding

4. HOW “SAVING TIME” MAKES A DIFFERENCE?

One might wonder whether bring the “saving time” emphasis to the frontbench in the
definition of visualization has a different implication from those existing definitions
given in Section 1. There are indeed some fundamental differences.

4.1 Measurement
Firstly, time is much easier to measure and quantify than insight, knowledge or
cognitive load, especially in the case of analytical tasks. In many ways, time may also
be easier to measure than information, that is, the quantitative measures used in

The first draft of this article was completed on 20 February 2013

What is Visualization Really for? Page 9

information theory. While the measurement about insight or cognitive load may be
undertaken in a laboratory condition, it is usually far too intrusive for a practical
environment. Such a measurement would be uncertain as the measurement
introduces a significant amount of artefacts and distortion to a normal cognitive
process of gaining insight.

4.2 Empirical Studies

Most empirical studies involved measurement of accuracy and response time. It is
comforting to know such measurements are not only meaningful, but also
fundamental. While we encourage and experiment with other studying methods, it is
important not to underestimate the measurement of time.
It is necessary to recognise the limitation of empirical studies in assessing “insight
gained”, especially when domain-specific knowledge is required. “Insight gained”
depends on data as well as existing knowledge of participants. When such knowledge
varies dramatically from one person to another, the study results have to be treated
with care. Hence empirical studies should focused on fundamental questions in
visualization, and have to minimise variables, especially those hard-to-observe and
hard-to-control variables such as a priori knowledge and insight to be gained.

4.3 Design Optimization

Measuring the time taken to perform a task can often be done seamlessly by a
system, subject to the necessary ethical consideration and user consensus. This
provides a metric for guiding the optimisation of the design of visual representations,
interaction methods, or visual analytics processes. In comparison with other metrics,
the time required to perform a task is undoubtedly the most important. It is easier to
measure, more objective, and more generic to all types of data, visual designs,
systems, tasks and users.

4.4 Theory of Visualization

The visualization community has not yet found a theory of visualization that most
would agree to be fundamental. The good news is that many researchers are inspired
to find such a theory, and some frameworks have been proposed. Any theory of
visualization should try to account for the impact of time required for performing or
accomplishing visualization tasks.

4.5 A Practical Wisdom

Most visualization researchers have had some experience of engaging with scientists
or scholars in different disciplines, or potential users from industrial or governmental
organizations. Many of us had encountered difficulties in persuading potential
collaborators about the merits of using visualization, or the need for developing
advanced visual designs and visualization systems. After demonstrating some
visualization techniques, typically conversations between a visualization researcher
and a potential user might flow like that:
Potential user (engagingly): These pictures are very pretty. We are interested in
having such techniques. I wonder how I can justify the costs for developing the
system that you proposed.
Visualization researcher (enthusiastically): As you can see from the demo,
visualization enables you to gain new insights from the data, this very much
outweighs the development costs.

The first draft of this article was completed on 20 February 2013

Page 10 Min Chen, Luciano Floridi, and Rita Borgo

Potential user (doubtfully): Really, what kind of insights are we talking about?
Visualization researcher (anxiously): Patterns. (Pause, trying to recollect some
definitions of visualization.) Interesting patterns, such as various anomalies,
complex associations, warning signs, and potential risks.
Potential user (hopefully but cautiously): Can those pictures tell me all these
automatically?
Visualization researcher (truthfully but uneasily): Not quite automatically. The
mapping from data to visual representations will enable you see these patterns
more easily and help you to make decisions.
Potential user (disappointedly): I can understand my data with no problem. I
could not imagine how these pictures can help me make better decisions.
After a while, some of us learned a wisdom, i.e., never suggesting to potential
collaborators that visualization could offer them insight. It is much better to state
that visualization could save their time. As Sections 2 and 3 have shown,
visualization can indeed save time.
“Gaining insight” has been an elusive purpose of visualization for several decades.
It is perhaps the time to invigorate visualization as a scientific discipline by shining
the spotlight on a more concrete purpose, that is, to save the time required for
accomplish a visualization task.

REFERENCES
S. Bateman, R.L. Mandryk, C. Gutwin, A. Genest, D. McDine and C. Brooks. 2010. “Useful junk?: the
effects of visual embellishment on comprehension and memorability of charts”. In Proc. ACM CHI,
2573–2582.
R. Borgo, A. Abdul-Rahman, F. Mohamed, P.W. Grant, I. Reppa, L. Floridi, and M. Chen. 2012. “An
empirical study on using visual embellishments in visualization”. IEEE Transactions on Visualization
and Computer Graphics, 18, 12, 2759-2768.
S. Card, J. Mackinlay, and B. Shneiderman. 1999. Readings in Information Visualization: Using Vision to
Think, Morgan Kaufmann Publishers.
S. Card. 2007. “Information visualization”. In A. Sears and J.A. Jacko (Eds.), The Human-Computer
Interaction Handbook, Lawrence Erlbaum Assoc Inc, 2007.
M. Chen and H. Jänicke. 2010. “An information-theoretic framework for visualization”. IEEE Transactions
on Visualization and Computer Graphics, 16, 6, 1206–1215.
W.S. Cleveland. 1985. The Elements of Graphic Data. Brooks Cole.
R.A. Earnshaw and N Wiseman. 1992. “An Introduction to Scientific Visualization”. In Scientific
Visualization, Techniques and Applications, K.W. Brodlie et al. (Eds), Springer-Verlag.
S. Few. 2009. Now You See It, Analytics Press.
S. Few. 2011a. The Chartjunk Debate: A Close Examination of Recent Findings.
http://www.perceptualedge.com/articles/visual_business_intelligence/the_chartjunk_debate.pdf
S. Few. 2011b. Benefitting InfoVis with Visual Difficulties? Provocation Without a Cause.
http://www.perceptualedge.com/articles/visual_business_intelligence/visual_difficulties.pdf
R.M. Friedhoff and T. Kiley. 1990. “The eye of the beholder”. Computer Graphics World, 13, 8, 46.
M. Friendly. 2007. “A brief history of data visualization”. In Handbook of Data Visualization, Springer, 15-
56.
H.G. Funkhouser. 1936. “A Note on a tenth century graph”. Osiris, 1, 260–262.
S.R. Gomez, R. Jianu, C. Ziemkiewicz, H. Guo and D.H. Laidlaw. 2012, “Different strokes for different
folks: visual presentation design between disciplines”. IEEE Transactions on Visualization and
Computer Graphics, 18, 12, 2411-2420.
M.A. Hearst. 2009. Search User Interfaces, Cambridge University Press.
N. Holmes. 1984. Designer’s Guide to Creating Charts and Diagrams, Watson-Guptill Publications.
J. Hullman, E. Adar, and P. Shah. 2011. “Benefitting infovis with visual difficulties”. IEEE Transactions
on Visualization and Computer Graphics, 17, 12, 2213 –2222.
B.H. McCormick, T.A. DeFanti and M.D. Brown (Eds.). 1987. Visualization in Scientific Computing, ACM

G.S. Owen. 1999. HyperVis  Teaching Scientific Visualization Using Hypermedia. ACM SIGGRAPH
SIGGRAPH Computer Graphics, 21, 6.

Education Committee. http://www.siggraph.org/education/materials/HyperVis/hypervis.htm

The first draft of this article was completed on 20 February 2013

What is Visualization Really for? Page 11

H.C. Purchase, N. Andrienko, T.J. Jankun-Kelly, and M. Ward. 2008. “Theoretical foundations of
information visualization. In Information Visualization: Human-Centered Issues and Perspectives, A.
Kerren et al. (Eds.) Lecture Notes In Computer Science, Vol. 4950. Springer-Verlag, 46-64.
H.Senay and E. Ignatius. 1994. “A knowledge-based system for visualization design”, IEEE Computer
Graphics and Applications, 14, 6, 36-47.
R. Spence. 2007. Information Visualization: Design for Interaction, Pearson.
E. Styles. 2006. The Psychology of Attention, Psychology Press.
A.C. Telea. 2008. Data Visualization, Principles and Practice, A K Peters.
J.J. Thomas and K.A. Cook (Eds.). 2005. Illuminating the Path: The Research and Development Agenda for
Visual Analytics, IEEE.
E.R. Tufte. 2001. The Visual Display of Quantitative Information, Graphics Press.
C. Ware. 2004. Information Visualization: Perception for Design, Morgan Kaufmann.

The first draft of this article was completed on 20 February 2013