Shuichiro Ogawa
日本語

Notes · updated 2026-09-22

What Data Visualization and Data Art Share, and Where They Differ

The same line chart is called a chart in a newspaper and data art on a gallery wall. The two share their material, their visual variables, their makers, and even their venues, so the medium and the tools cannot separate them.

Contents (9)
  1. The material, the tools, and the makers are the same
  2. ”Visualization is accurate, art is beautiful” does not survive the evidence
  3. The difference is whether the figure demands a single reading
  4. Not a binary, but a continuum between two poles
  5. Can the distinction be settled objectively?
  6. What can be settled objectively is which one a work was treated as
  7. Telling them apart across three layers
  8. Where the distinction is needed, and where it is not
  9. What is still unresolved

Take a single line chart. Print it in a newspaper and it is called a chart. Hang it on a gallery wall and it is called data art.

The name changes what counts as praise. The newspaper chart is called clear and accurate, and a misreading is called misleading. The gallery piece is called beautiful and thought-provoking, and being hard to read is not held against it.

The shape of the figure is the same. Yet what counts as good and what counts as failure swap places. What is shared, what differs, and can the difference be settled from the outside?

The material, the tools, and the makers are the same

Start with the parts that cannot be told apart. Data visualization and data art overlap on at least four fronts.

First, the material. Both work from collections of numbers. Viégas and Wattenberg defined artistic data visualization as visualization of data done by artists with the intent of making art, and added a second condition: the work must be based on actual data, rather than the metaphors or surface appearance of visualization. A piece that does not use data falls outside that definition from the start.

Second, the tools for mapping data onto vision. Bertin treated graphics as a sign system and laid out seven visual variables: position, shape, orientation, hue, texture, value, and size. Visualization and data art both turn data into vision using that same set of variables.

Third, the makers and the tools. The same people make both, often with the same drawing libraries. The two are not separate professions.

Fourth, the venues. IEEE VIS, the main international conference on visualization, runs VISAP, an art program whose juried works are exhibited alongside the conference.

The history is shared as well. ISOTYPE, built by Otto Neurath and Gerd Arntz in the 1920s, used pictures to make statistics legible for public education. Jansen traces that lineage and records how ISOTYPE lost influence in statistical graphics in the later twentieth century while persisting in art and design. The apparatus of the statistical chart was handed over to art intact.

”Visualization is accurate, art is beautiful” does not survive the evidence

Set aside, for a moment, the usual split. Visualization belongs to accuracy, art belongs to beauty. That split breaks on both sides.

Start with visualization. Lisnic and colleagues analyzed 10,000 Twitter posts with COVID-19 visualizations and found 84% of posts that drew an interpretation were misleading. Yet only 11% violated common design guidelines, and fewer than 1% of charts were misread because of visual tricks. The dominant cause was errors of reasoning: cherry-picked time windows, arbitrary thresholds, causation read out of correlation. Following the guidelines does not make a chart truthful.

Visualization also carries techniques of persuasion. Hullman and Diakopoulos showed how adding and omitting at the layers of data, representation, annotation, and interaction privileges particular interpretations. They call this visualization rhetoric. Visualization is not neutral transmission.

The art side is no freer of truth claims. Offenhuber describes reading phenomena and the process of data generation from material traces and environmental indicators, a practice he calls autographic visualization, and argues that data art can constitute evidence. Drucker calls for rebuilding the conventions of graphical display so that interpretation is built into the figure itself. Manovich framed data visualization art as the anti-sublime: mapping phenomena beyond human perceptual scale into representations at human scale, and he flagged the arbitrariness of that mapping as the dark side of the form.

Decoration follows the same pattern. Tufte set the elimination of chartjunk and the reduction of non-data ink as a norm. Bateman and colleagues tested that norm and found decorated charts no worse on comprehension and significantly better on recall two to three weeks later. Decoration is not automatically the enemy.

The difference is whether the figure demands a single reading

So where does the difference lie? Both the theory and the empirical work point not at appearance but at what the figure asks of its reader.

Definitions in visualization put the purpose on the reader’s task. Card, Mackinlay, and Shneiderman define it as the use of computer-supported, interactive, visual representations of abstract data to amplify cognition. Munzner’s textbook definition describes computer-based visualization systems that provide visual representations of datasets designed to help people carry out tasks more effectively, and limits the fit to cases where human capabilities are augmented rather than replaced. Mackinlay translated quality into two criteria: whether every piece of the data table is represented, and whether the mapping is faster to interpret, conveys more distinctions, or leads to fewer errors than an alternative. Visualization is measured by speed, accuracy, and error.

On the art side, the number of readings changes. Pousman, Stasko, and Mateas surveyed boundary cases of information visualization, including ambient, social, and artistic displays, and observed that core systems assume a single correct reading of the visual output, whereas boundary cases can admit multiple correct interpretations of the underlying data. Ambiguity is treated as design space rather than defect.

Lan and colleagues classified 220 data art works from VISAP 2013 to 2023 and interviewed twelve data artists. The stated intents were experiment (49), inform (39), engage (36), criticize (19), and witness (6). In the interviews, the artists described the difference from visualization in terms of how information is conveyed. The authors characterize artists as accepting subjectivity, particularity, and ambiguity over objectivity.

Put plainly: visualization casts the reader as a worker and treats a misreading as failure. Art casts the reader as an interpreter and does not treat an excess of readings as failure. The yardsticks for success are simply different.

Not a binary, but a continuum between two poles

That difference does not work as a sorting rule, because works carry both sets of properties.

Kosara argued that classification in visualization leaned on technical criteria and dropped artistic work, and proposed treating artistic visualization and pragmatic visualization as a spectrum rather than two kinds. Lau and Vande Moere defined information aesthetics as the conceptual link between information visualization and visualization art, placing it as the middle term that connects rather than separates them. The researchers themselves assume a middle ground exists.

In practice, Ludwig and colleagues reviewed curating exhibitions of artistic data visualization and recorded that such work is usually shown alongside a data visualization conference, with group exhibitions independent of academic events remaining rare. In the exhibition hall, the two sit side by side.

Can the distinction be settled objectively?

This is the question at the center. Can an outside observer decide whether a given figure is visualization or art? Academic answers divide into three.

One position holds that properties of the object decide. Beardsley defined art as something produced with the intention of giving it the capacity to satisfy aesthetic interest. On this view the criterion lives in the object and does not vary with the judge. According to the Stanford Encyclopedia of Philosophy, however, aesthetic definitions draw counterexamples in both directions: too narrow, excluding avant-garde works that place ordinary objects in galleries, and too wide, admitting cars and commercial design.

A second position holds that no single criterion exists. Weitz argued that the concept of art is open, so its necessary and sufficient conditions cannot be fixed in principle; each new work forces a revision of the concept’s extension. The source of that argument is Wittgenstein’s family resemblance. Gaut recast it as the cluster account: aesthetic properties, emotional expression, intellectual challenge, complex meaning, originality, high skill, and other criteria, none of them necessary, with a sufficient subset doing the work. On this view the judgment becomes a matter of degree rather than a binary.

A third position holds that institutions decide. Danto argued that what turns a perceptually indistinguishable object into a work is an interpretation backed by art theory and art history, and the institutional context that makes such interpretation possible. Dickie turned that into a definition: an artifact upon which a society or subculture has conferred the status of candidate for appreciation. Davies groups definitions of this type as proceduralist. Art status is still objective here, but the objectivity sits in the institution rather than in the object. The same item can therefore get different answers depending on which institution handles it.

Empirical work leans toward the third position. Pelowski and colleagues had 114 participants classify 140 images as art or not art. Even for Renaissance and Baroque paintings, only 72.6% of participants classified every image as art, and only 5.4% answered yes on every trial. Readymades were classified as art 47.8% of the time and photographs of everyday objects 14.9%. Judgments correlated with the beliefs participants held, such as the idea that a painting is automatically art, more than with properties of the images.

Hekkert and van Wieringen had 34 experts and 26 non-experts rate works and reported low agreement between raters, with experts not necessarily agreeing more than non-experts. [primary check pending] Tröndle and colleagues studied museum visitors and found that judgments of contemporary art were driven by prior knowledge, sociodemographic background, and emotional experience. Mikalonytė and Kneer ran preregistered experiments with 888 participants, manipulating intentional creation, aesthetic value, and institutional recognition, and found that none of the three was sufficient on its own.

Machine classification has been attempted as well. Brachmann and colleagues distinguished traditional paintings from non-art photographs with 93.0% accuracy using only two summary measures over CNN features, against a chance level of 0.513. The authors themselves limit the result: the classifier fails on modern and postmodern art. The cues that work for traditional painting are not a definition of art in general.

The conclusion is that no method settles the question objectively from properties of the object alone. The judgment is a function of the object, the judge, and the setting together.

What can be settled objectively is which one a work was treated as

Stopping at “undecidable” leaves nothing usable. So restate the question.

What can be recorded objectively is not which one a work is, but which one it was treated as. Who certified it, when, on what criteria, and as what: those are facts that persist.

On the visualization side, that certification is public. Reyes, as program chair, explained the position of the SIGGRAPH Art Papers program in a journal article, recording how works are selected and what standing the program has. For VISAP, Ludwig and colleagues wrote up the curatorial process as a practice report. There is a jury, a selection, an exhibition. The fact that a work was treated as art accumulates on the side of the procedure.

That lets the question be rewritten. Not “is this art” but “which institution certified this, and as what.”

Telling them apart across three layers

For practical use, ask the following three layers in order.

The layer of demands: what does the figure ask of its reader? Does it ask the reader to extract numbers? Are there axes, a legend, units, a source? If a reader gets it wrong, can that be called a mistake? If the answers are yes, the figure was made as visualization.

The layer of intent and venue: what was it made for, and where was it shown? A peer-reviewed visualization conference, a gallery, a juried competition? This is the layer where Viégas and Wattenberg placed their definition.

The layer of institutions: who certified it, and as what? Accession, exhibition, criticism, grants, publication. The artworld that Danto and Dickie describe lives on this layer.

Each layer has a failure mode. The layer of demands breaks when an accurately readable figure is accessioned as a work. The layer of intent breaks when the maker’s intent cannot be read from outside, or gets retold later. The layer of institutions breaks when institutions disagree, which they are free to do: a conference may treat a work as visualization while a museum treats it as art.

The final answer, then, is not “undecidable.” It is to decide which institution you are asking, and to check that institution’s criteria. Accounting, funding, peer review, and accession each hold an answer inside their own procedure.

Where the distinction is needed, and where it is not

Use the three layers where a decision has to be made. Which budget pays for it. What counts as success. Which contract it is made under. Because the yardsticks differ, deciding late means good work gets judged as failure.

In appreciation, no decision is needed. Whichever the figure is, the reading will not settle into one. As the artists interviewed by Lan and colleagues put it, holding several readings at once is what the work does.

What is still unresolved

Gaps confirmed during collection for this note.

  • No study was found that directly measures inter-rater agreement when visualizations and data art are judged side by side. Agreement has been measured for art in general (Pelowski et al., Hekkert and van Wieringen), but not for that contrast.
  • The review criteria and acceptance rates of VISAP and SIGGRAPH Art Papers were behind paywalls. The account of the institutional layer therefore rests on what is publicly documented.
  • Japanese-language aesthetics and design research on the art and design distinction was not searched for this pass.

References

  • Adajian, T. (2007, revised 2024). “The Definition of Art”. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/art-definition/ (accessed 2026-09-22)
  • Bateman, S., Mandryk, R. L., Gutwin, C., Genest, A., McDine, D., & Brooks, C. (2010). “Useful Junk? The Effects of Visual Embellishment on Comprehension and Memorability of Charts”. CHI ‘10, 2573-2582. https://doi.org/10.1145/1753326.1753716
  • Beardsley, M. C. (1983). “An Aesthetic Definition of Art”. In H. Curtler (ed.), What Is Art?, Haven Publications.
  • Bertin, J. (1967/1983/2010). Semiology of Graphics: Diagrams, Networks, Maps. Esri Press. ISBN 9781589482616.
  • Brachmann, A., Barth, E., & Redies, C. (2017). “Using CNN Features to Better Understand What Makes Visual Artworks Special”. Frontiers in Psychology, 8, 830. https://doi.org/10.3389/fpsyg.2017.00830
  • Card, S. K., Mackinlay, J. D., & Shneiderman, B. (eds.) (1999). Readings in Information Visualization: Using Vision to Think. Morgan Kaufmann. ISBN 9781558605336.
  • Danto, A. C. (1964). “The Artworld”. The Journal of Philosophy, 61(19), 571-584. https://doi.org/10.2307/2022937
  • Davies, S. (1991). Definitions of Art. Cornell University Press. https://doi.org/10.7591/9781501721182
  • Dickie, G. (1969). “Defining Art”. American Philosophical Quarterly, 6(3), 253-256. (Expanded in 1974 as Art and the Aesthetic: An Institutional Analysis, Cornell UP.) [primary check pending]
  • Drucker, J. (2011). “Humanities Approaches to Graphical Display”. Digital Humanities Quarterly, 5(1). https://dhq.digitalhumanities.org/vol/5/1/000091/000091.html
  • Gaut, B. (2000/2005). “The Cluster Account of Art Defended”. British Journal of Aesthetics, 45(3), 273-288. https://doi.org/10.1093/aesthj/ayi032
  • Hekkert, P., & van Wieringen, P. C. W. (1996). “Beauty in the Eye of Expert and Nonexpert Beholders: A Study in the Appraisal of Art”. American Journal of Psychology, 109(3), 389-407. https://doi.org/10.2307/1423013
  • Hullman, J., & Diakopoulos, N. (2011). “Visualization Rhetoric: Framing Effects in Narrative Visualization”. IEEE TVCG, 17(12), 2231-2240. https://doi.org/10.1109/TVCG.2011.255
  • Jansen, W. (2009). “Neurath, Arntz and ISOTYPE: The Legacy in Art, Design and Statistics”. Journal of Design History, 22(3), 227-242. https://doi.org/10.1093/jdh/epp015
  • Kosara, R. (2007). “Visualization Criticism - The Missing Link Between Information Visualization and Art”. Proc. IV ‘07, 631-636. https://doi.org/10.1109/IV.2007.130
  • Lan, X., Wang, Y., Peng, L., & Ma, X. (2025). “More Than Beautiful: Exploring Design Features, Practical Perspectives, and Implications of Artistic Data Visualization”. IEEE PacificVis 2025, 329-339. https://doi.org/10.1109/PacificVis64226.2025.00039
  • Lau, A., & Vande Moere, A. (2007). “Towards a Model of Information Aesthetics in Information Visualization”. Proc. IV ‘07, 87-92. https://doi.org/10.1109/IV.2007.114
  • Lisnic, M., Polychronis, C., Lex, A., & Kogan, M. (2023). “Misleading Beyond Visual Tricks: How People Actually Lie with Charts”. CHI ‘23, Article 817. https://doi.org/10.1145/3544548.3580910
  • Ludwig, L., Castro, B., & Kosminsky, D. (2024). “Numerical Existence: Reflections on Curating Artistic Data Visualization Exhibitions”. 2024 IEEE VIS Arts Program (VISAP), 69-77. https://doi.org/10.1109/VISAP64569.2024.00010
  • Mackinlay, J. D. (1986). “Automating the Design of Graphical Presentations of Relational Information”. ACM Transactions on Graphics, 5(2), 110-141. https://doi.org/10.1145/22949.22950
  • Manovich, L. (2002). “Data Visualization as New Abstraction and Anti-Sublime”. manovich.net. https://manovich.net/index.php/projects/data-visualisation-as-new-abstraction-and-anti-sublime
  • Mikalonytė, E. S., & Kneer, M. (2024). “The folk concept of art”. Synthese, 205(1), Article 2. https://doi.org/10.1007/s11229-024-04812-8
  • Munzner, T. (2014). Visualization Analysis and Design. CRC Press / A K Peters. ISBN 9781466508910. https://www.cs.ubc.ca/~tmm/vadbook/
  • Offenhuber, D. (2019). “Data by Proxy - Material Traces as Autographic Visualizations”. IEEE TVCG, 26(1), 98-108. https://doi.org/10.1109/TVCG.2019.2934788
  • Pelowski, M., Gerger, G., Chetouani, Y., Markey, P. S., & Leder, H. (2017). “But Is It really Art? The Classification of Images as ‘Art’/‘Not Art’ and Correlation with Appraisal and Viewer Interpersonal Differences”. Frontiers in Psychology, 8, 1729. https://doi.org/10.3389/fpsyg.2017.01729
  • Pousman, Z., Stasko, J., & Mateas, M. (2007). “Casual Information Visualization: Depictions of Data in Everyday Life”. IEEE TVCG, 13(6), 1145-1152. https://doi.org/10.1109/TVCG.2007.70541
  • Reyes, E. (2019). “Art Papers Jury: Introducing the SIGGRAPH 2019 Art Papers”. Leonardo, 52(4), 346-348. https://doi.org/10.1162/LEON_a_01775
  • Tröndle, M., Kirchberg, V., & Tschacher, W. (2014). “Is This Art? An Experimental Study on Visitors’ Judgement of Contemporary Art”. Cultural Sociology, 8(3), 310-332. https://doi.org/10.1177/1749975513507243
  • Tufte, E. R. (1983/2001). The Visual Display of Quantitative Information. Graphics Press. ISBN 096139210X. https://www.edwardtufte.com/books/
  • Viégas, F. B., & Wattenberg, M. (2007). “Artistic Data Visualization: Beyond Visual Analytics”. OCSC 2007, LNCS 4564, 182-191. https://doi.org/10.1007/978-3-540-73257-0_21
  • Weitz, M. (1956). “The Role of Theory in Aesthetics”. Journal of Aesthetics and Art Criticism, 15(1), 27-35. https://doi.org/10.2307/427491
  • Wittgenstein, L. (1953/2009). Philosophical Investigations (4th ed.). Wiley-Blackwell. ISBN 9781405159289.

Author: Shuichiro Ogawa (Design Researcher / Consultant) About me →