This essay looks into the matters of neutrality and transparency in data visualisation design. More specifically, it disaggregates the several "data strata" involved in the production and consumption of data visualisation, of which, amongst the numeric and visual, the designer is also one; and subsequently proposes how each stratum should contribute to seeing data and its visualisations as subjective rather than objective practices. Finally, it touches upon the accountability that the designer and responsibility that the reader hold when engaging with a piece of data visualisation design.

Data visualisation: an in-between
Data Visualisation is often described as incorporating elements from many disciplines: map design principles from cartography, conventions of displaying numerical data in charts from statistics, best practices for the use of type, layout, and colour from graphic design, and principles of writing and storytelling from journalism, amongst others (Cairo 2013: 23). While this is advantageous in that it allows a lot of stretching and morphing, it also means that data visualisation is embedded with certain predetermined expectations. Predominantly, these have to do with the common perception for data visualisations to be neutral (in the data¹) and transparent (in the visuals²). In the recently published book ‘Data visualization in society’ (2020) Kennedy, Weber, and Engebretsen identify these two dimensions as the numeric stratum and the visual stratum (172).
Regarding the numeric stratum, questions of how and why data is collected can be taken back to the earliest note-taking practices (Gitelman 2013); however, the increasing relevance of data science and data visualisation has prompted the interrogation of what we even understand data itself to be, as well as how we formulate claims of truth (Braun 2017: 9). In the fields of statistics and data science, data is often described as being in ‘raw’ state – in view of the fact that it represents collected phenomena with a disembodied nature from its human agent. Moreover, this data is often numeric, which in turn questions our perception of whether numbers are neutral. The 18th-century scientific revolution’s centrality in mathematical truth meant that facts were universal and thus eternally preexisting (Rock 2013: 48), making numbers ‘historically trusted because they appear universal, impersonal, and neutral’ (Kennedy, Weber, Engebretsen 2020: 172). This perception was strengthened with the rise of mechanical objectivity in photography, which as Gittelman points out ’emerged as a dominant ideal in the sciences only in the middle of the nineteenth century’ – with the first photographic processes, ‘observers were struck by the apparent displacement of human agency in the production of life-like images’ (2013: 5). This leant photography – a type of visual representation – a sense of neutrality: mechanical photography was the unbiased agent that human-made painting wasn’t. This came to be questioned, later on, but is still embedded both in the perception of numbers as objective and in the visual nature of data visualisations. Borrowing Gitelman’s expression (2013), there is no such thing as raw data: it’s an oxymoron.
As for the visual stratum, there is first the form/content discourse in graphic design, questioning the role of how ‘visible’ or transparent the visual in graphic representations should be. The epitome of this is Beatrice Warde’s 1930 advocation for typography to be transparent, creating no obstruction to its content, text (the metaphor being a crystal goblet not fogging the quality of a good wine). Extrapolated to the field of data visualisation, this is the equivalent of what Edward Tufte presents in his influential book ‘The Visual Display of Quantitative Information’ (2001): transparency for him is the belief that visualisations should be an objective representation of data that do not obstruct or skew it. But as Michael Rock points out, the focus has been erroneously been on the goblet, when, in reality, it’s all about the wine (2013: 92). Second, as pertaining to the visual realm, graphics have much to do with sight, and in that sense with writing more than with speaking. It is precisely ‘the visual nature of writing [that] leads to ideas of objectivity that are impossible in oral culture’, as writing ‘separates the knower from the known’ (Rettberg 2020: 37). Calvert too identifies that the oral qualities of language have long been ‘held to be aesthetically central to literature (…) while the visual is largely regarded as aesthetically and semantically irrelevant’ (2012: 314). As graphic representations, data visualisations are embedded with this distrust of writing and graphicality, one the one hand, and with the ideal of transparency of form not obfuscating content, on the other.
Whether it’s the visual that allows for an analytical and objective stance or the quantitative nature of data visualizations that ‘lend them this sense of authority’ (Rettberg 2020: 39), as Goatley evidences much of the discussion of the previous paragraphs ‘is not a question of aesthetics, but of ontology: graphics do not transparently reveal data – they create a new subjective interpretation of it’ (2019: 25). In fact, and in the same way that data is different to that which it measures, any graphic representation of information is phenomenologically distinct to the information itself.
Designer stratum
In an attempt to disaggregate what neutrality and transparency mean in data visualisation, the arguments presented above focused on analysing the numeric and visual strata. However, as Alberto Cairo identifies, in turning data into structured information there’s the communicator that gives ‘shape to data’ (2013: 16). By designing visualisations, the designer is an agent in the whole process, and, inevitably, another piece of data. Just as numbers and data don’t just exist, data visualisations don’t, either: they are created for specific reasons, with a purpose, and by someone. It is relevant to bring this figure to the surface as it is the agent who is precisely interpreting the numeric stratum into the visual stratum, and, as such, at the centre of these discourses – a designer stratum itself. Their role includes either all or part of the thinking, collecting, documenting, mapping, analysing, reflecting, translating, synthesising, and concluding around the data and the graphics.
Designer as author
Conceiving the designer as an agent raises questions of authority. In graphic design discourse, as Michael Rock points out, authorship allows ‘understanding the design process in a profession traditionally associated more with the communication than the origination of messages’ (2013: 46). Moreover, it could be argued that the communication of messages is itself an act of origination, as it implies a set of decisions determined by the communicator. Interestingly, a parallel of the convergence of science and art in data visualisation can be found in the notion of authorship. Rock identifies that until fairly recently, authorless texts in literature had more claims to authenticity than authored ones. He goes on to say that, on the other hand, scientific texts, ‘at least through the Renaissance, demanded an author’s name as validation’ (Rock 2013: 47). This is because science was based in subjective invention and the authority of the scientist rather than on an objective truth. However, with the Scientific Revolution’s centrality in mathematical truth the situation reverses: literature was authored and science the product of anonymous objectivity. The ownership of text and authority that this granted to authors in literature ‘fueled much of the 20th century’s obsession with authorship’ (Rock 2013: 48). Being associated with writing and of a graphic nature, the graphic arts adopted many of these discourses – and so did graphic design.
Data visualisation designers have thus inherited both the authorlessness of the scientist and that of the graphic artist. This lack of authority arguably contributes to the idea of data visualisation as an objective truth. And even though the notion of authorship is an important one to take into account, as it brings to the forefront the figure of the someone behind a piece, it is rather the understanding of how this someone shapes the piece that it’s most relevant to analyse. Whether they are the generators or facilitators of content, it is precisely how the designer’s view of the world shapes their designs that can, in my view, be claimed as a mode of authorship. The designer is a communicator that has agency in the message they are translating into visual form.
Designer as data
What does the designer stratum bring to the data strata? As creative practitioners, we bring ourselves to everything we do: as Friedman states, the designer is not ‘the centre of the design process itself’ but ‘the psychological centre of his or her personal perceptual process’ (2001: 39). We are also ‘intimately, physically connected to the work we produce, and it is inevitable that our work bears our stamp’ (Rock 2013: 95). As such, this personal perceptual process will also affect the neutrality and transparency in data visualisation. As Gitelman notes, ‘different visualizations are differently effective, well or poorly designed, and all data sets can be multiply visualized’ (2013: 12). Individual designs, as well as ideas about designing, are related to wider sets of values.
As Kennedy and Engebretsen state, ‘the potential meanings carried by semiotic resources are dependent on both cultural conventions and the particular situations of use, including the background and motivations of the human participants’ (2020: 25); similarly, Jill Simpson identifies data visualisations as having a ‘situated nature’ (2020: 162). Just as researchers need to be reflexively aware of how their communities of practice act upon them and how they act upon those (Crouch, Pearce 2012: 42), designers too should adopt a reflexive position in regards to their graphic production. All of these influencing bodies shape the design in ways that might be invisible to the audience, but that carry with them the imprint of the designer.
Recognising that we are all the products of our individual lived experiences (without disregarding the agency that we have in our positioning within the wider human congregate), is it worth considering a softer approach to the perception of data visualisation designers? Understanding them more as ‘material-semiotic storytellers’ (Ward 2008: 5) than as truth-revealers, the designer as ‘an expert endowed with the power to create form, solve problems, pass judgment, and confer meaning’ (Lupton, Abbot-Miller 1996: 70)? This reflexive position should include accountability too, and in order to engender trust data visualisers may accordingly need to be open about their choices. One of the principles of the ‘data feminism’ movement by Lauren Klein and Catherine D’Ignazio, Embrace Pluralism, proposes how self-disclosure and an embrace of pluralism can expose the decisions that contribute to the creation of data visualisations, in this way shifting from the current emphasis on objectivity in favour of designs that facilitate pathways to multiple truths (2020: 125). Perhaps the question is not one of truthfulness but of reliability, so that designers can reject the discourse of transparency and instead actively adopt a reflexive position where they are able to understand the relationship between their own and others’ ideological assumptions in the acts of communication and meaning-making.
The reader
Finally enveloping the data strata is the layer of which this production of meaning-making devices is reason of: the reader³. It is important to place greater emphasis on the responsibility that the reader has as a consumer: rather than demanding for transparency on the side of the data visualiser, they should acknowledge that the choice to engage or not with a data visualisation design is ultimately down to them, and that they must do so knowing that they are consuming not only the product of someone else’s mind, but also that of a broader discourse-at-large. As Rock notes, any translation ‘reflects both the characters of the original and the spirit of the contemporary as well as the individuality of the translator’ (2013: 54) – at its best, graphic design is ‘merely reflecting the sentiments that culture at large is already feeling’ (Stinson 2018: np). It’s the troublesome relationship with data, numbers, graphicality, and authorship discussed previously that seem to set the designer aside of other producers of knowledge, arguing that as pieces of knowledge-generation visualisations should be neutral. But they are – just as any other occurrence of the human-made – another instance of data as imperfect, subjective, and complex (Lupi 2017). The challenge as readers is in understanding the multiplicity of methods that can comprise a design language, as well as in deciding ‘how to express the nuances of our personal identities in our design choices’ (Rawsthorn 2018: 75). As readers, too, we can become reflexive of our choices.
Data strata
Rather than understanding data visualisation as an aggregate of a numeric and visual stratum that are allegedly neutral and transparent and become even more so when merged, the focus should be in considering it a distinct entity that rephrases, reinterprets, and presents content anew. Neither data, its visualisation, or its visualiser ‘represent an a priori truth, but offer constructed, fallible, and subjective views of the world’ (Goatley 2019: 25). Even though data visualisation has suffered from suspicion in regards to claims of truth because of its inherited history, it is increasingly a way in which we mediate the information we produce. As human beings, all we ever do is to try and make sense of things, and as such it’s pertinent to now ‘lift that ban of suspicion and engage the full potential of visuality to produce and encode knowledge as interpretation’ (Drucker 2014: 11). Should there be greater emphasis in advocating for more ‘data literacy’? As Johanna Drucker beautifully proposes in her book Graphesis, it is crucial that we recognise how much of our lives is mediated and shaped by data visualisation, and this calls for greater comprehension of the graphic realm we constantly engage with. Instead of expecting production and consumption of neutrality and transparency in data visualisation, there should be nurturing of more crucial thinking to instruct us all to deal with the capacity to critically engage with the graphic devices that surround us.
The image at the beginning of this article is a proposed diagram that maps all of the elements involved in data visualisation outlined in the article. It uses the metaphor of geological ‘strata’ to evidence all the layers and to reference data visualisation’s connection with map-making and diagrammatic structures. It is based on Alberto Cairo’s diagram From reality to people’s brains (2013: 16, Figure 1.8), and on early diagrams of geological layers on Earth.


Notes
- Inasmuch as they are representing data, or information – that is, that the data used is itself of a neutral, unbiased, and objective nature.
- Given that they are of a graphic nature (in visually mapping, rendering, displaying, understanding, and reasoning information) – that is, that the graphic devices employed should not obfuscate the content they display.
- Recognising the dual role of data visualisation as a means of exploring data and also explaining it, this figure can also be considered to be the designer or even the data itself.
References
Braun, S. (Ed.) (2017). Data Visualisation for Sucess. Victoria The Images Publishing Group.
Cairo, A. (2013). The functional art. San Francisco: New Riders.
Calvert, S. (2012), ‘Materia Prima, text-as-image’, Journal of Writing in Creative Practice 4: 3, pp. 309–328, doi: 10.1386/jwcp.4.3.309_1
Crouch, C., Pearce, J. (2012). Doing research in design. ProQuest Ebook Central. Created from UAL on 2021–01–10. Bloomsbury Publishing Plc.
Drucker, J. (2014). Graphesis. Harvard: Harvard University Press.
D’Ignazio, C., Klein, L. (2020). Data Feminism. Cambrdige, Massachhusetts: The MIT Press.
Friedman, K. (2001). Creating design knowledge: From research into practice. IDATER 2000: International Conference on Design and Technology Educational Research and Curriculum Development.
Gitelman, L. (Ed.) (2013). "Raw data" is an oxymoron. Cambdridge: MIT Press. Goatley, W. (2019). Critical Data Aesthetics: Towards a Critically Reflexive Practice of Data Aestheticisation. Doctoral thesis, University of Sussex.
Kennedy, H., Weber, W., & Engebretsen, M. (2020). Data visualization and transparency in the news. In Kennedy H. & Engebretsen M. (Eds.), Data Visualization in Society (pp. 169–186). Amsterdam: Amsterdam University Press. doi:10.2307/j.ctvzgb8c7.17
Lupi, G. (2017, January 30). Data humanism: The revolutionary future of data visualization. Print Magazine. Retrieved from http://www.printmag. com/information-design/data-humanism-future-of-data-visualization.
Lupton, E., Abbot-Miller, J. (1996). Design Writing Research. London Phaidon.
Rawsthorn, A. (2018). Design as an attitude. Zürich: JRP Ringier.
Rettberg, J. (2020). Ways of knowing with data visualizations. In Engebretsen M. & Kennedy H. (Eds.), Data Visualization in Society (pp. 35–48). Amsterdam: Amsterdam University Press. doi:10.2307/j.ctvzgb8c7.8
Rock, M. (Ed.) (2013). Multiple Signatures: on designers, authors, readers and users. New York: Rizzoli.
Stinson, L. (2018) Is Millennial Minimalism On Its Way Out?, AIGA Eye on Design, accessed Jan 25th 2021.
Ward, M., Wilkie, A. (2008). Made in Criticalland: Designing Matters of Concern. Networks of Design: Design History Society Conference.