Centre for Health, Technologies and Social Practice

Big data and their visualisation: the ideological work of data visualisation conventions

By Rosemary Lucy Hill

(This short article is adapted from The Work that Visualisation Conventions Do by Helen Kennedy, Rosemary Lucy Hill, Giorgia Aiello and Will Allen)1

In our digital world, data are becoming increasingly valued and relied upon as tools in decision making, and as explainers of the world and societies. The idea that ‘with enough data, the numbers speak for themselves’2 is very powerful though not uncritiqued.3, 4 Data are most often understood through their representation in visual forms – some familiar such as graphs, bar charts, pie charts, and some less so, like stream graphs, isotype charts and Sankey diagrams (read more about these chart types at seeingdata.org).  I have been working on a project called Seeing Data which aims to understand how people (non-specialists) engage with these data visualisations. As well as speaking to many non-specialists, we also spoke to data visualisers about their practice and we analysed a number of visualisations.

When we analysed the visualisations we found that the conventions of the graphical representation of data do ideological work: they make visualisations seem objective, as if they are neutral windows onto factual data. They do this in a number of ways, and here I’m going to explain two of these conventions: two-dimensional viewpoints and geometric shapes and lines.

2011 Census Map Analysis – Percentage of Respondents Born in the UK (Office for National Statistics)

2011 Census Map Analysis – Percentage of Respondents Born in the UK (Office for National Statistics)

Two-dimensional viewpoints

Visualisations often use two-dimensional viewpoints, either through a top-down view, as in maps and pie charts, or a front-on view, as in graphs that use an x/y axis. For example in ONS’s 2011 Census Map Analysis, we hover above England and Wales and look at these countries from a two-dimensional perspective, making it possible to see the whole of these two nations as if we were everywhere at once. Although there are good reasons for using 2D perspectives rather than 3D viewpoints (it is more difficult to read the values from a 3D graph; distortions of perspective are avoided), using two-dimensional, front-on or top-down views, presents a different kind of perspective, but one that isn’t acknowledged. Think, for example, of the differences between Mercator and Peters map projections: whole continents shift in shape and size in the two different projections. But a view without apparent perspective is still a view from somewhere, as Haraway5 and others have argued; usually the view of the privileged and powerful.

 

Shapes and lines: creating order

The Global Flow of People (Nikola Sander, Guy J. Abel & Ramon Bauer, Wittgenstein Centre for Demography and Global Human Capital)

The Global Flow of People (Nikola Sander, Guy J. Abel & Ramon Bauer, Wittgenstein Centre for Demography and Global Human Capital)

Simple geometric shapes and lines are widely used in data visualisations. Rectangles and circles appear in bar and pie charts and also as markers and structuring devices on other types of visualisation. This privileging of repetitive geometric shapes creates a feeling of order, helping us to regulate expectations and ease our understanding.6 But abstraction is also associated with the universal. This association emphasises a belief in the ability of geometric shapes to offer the same meaning to all viewers and to present data which are universally true – another ‘god-trick’.5, p. 189 Furthermore, the use of geometric shapes  to represent many types of things, e.g. human beings, share prices, unemployment figures,7 is a reductive process, which simplifies data by highlighting only some characteristics and not others.

This is apparent in The Global Flow of People, a chord diagram showing global migration patterns. Individual people (indicated by the precise values shown when the ribbons are hovered over) seem to move from one country to another in smooth arcs. This suggests a simple journey, a straightforward transition from one nation-state to another. However, for many migrants the process of moving involves transiting through other countries, returns to the country of origin, and substantial struggles to gain recognition as migrants.8 The use of simple lines and shapes in the chart means that it is not possible to understand how different categories of migrants are grouped, which groups are being counted, and which are excluded. The Global Flow of People presents a particular version of human migration because visualisation designers are constrained by conventions.

These conventions are closely tied to the Enlightenment project of attempting to rationally understand the external world as a knowable object.7, 9 The conventions of data visualisation therefore have links to European imperial capitalist expansionism and the oppression of some groups (indigenous populations, women). The assertion of the neutrality and objectivity of the data displayed within visualisations therefore does ideological work: it presents a particular viewpoint as the whole truth and naturalises the dominance of those who have the power to present data. Interrogating the work that visualisation conventions do therefore helps us to understand how data visualisations are imbued with particular qualities and how power operates in and through them. What we need to do next is to ensure that data and data visualisations are not only the preserve of already privileged groups.10

 

You can read more about data visualisation conventions in the article I wrote with Helen Kennedy, Giorgia Aiello and Will Allen, and published in the journal Information, Communication and Society: The work that visualisation conventions do. You can read more about the Seeing Data project at seeingdata.org.

 

References

  1. Kennedy H, Hill RL, Aiello G and Allen W. The work that visualisation conventions do. Information, Communication & Society. 2016: 1-21.
  2. Anderson C. The end of theory. Wired. 2008.
  3. Bowker GC. Memory practices in the sciences. Cambridge, Mass. ; London: MIT, 2005.
  4. Gitelman L and Jackson V. Introduction. In: Gitelman L, (ed.). Raw Data is an Oxymoron. Cambridge, MA: MIT Press, 2013, p. 1-14.
  5. Haraway D. Situated knowledges: The science question in feminism and the privilege of partial perspective. Feminist studies. 1988: 575-99.
  6. Arnheim R. Perceptual abstraction and art. Psychological Review. 1947; 54: 66.
  7. Manovich L. What is visualisation? Visual Studies. 2011; 26: 36-49.
  8. Ahrens J, Kelly M and Liempt I. Free movement? The onward migration of EU citizens born in Somalia, Iran, and Nigeria. Population, Space and Place. 2014.
  9. Friendly M. A brief history of data visualization. In: Chen C-h, Hardle W and Unwin A, (eds.). Handbook of data visualization. Berlin: Springer, 2008, p. 15-56.
  10. Hill RL, Kennedy H and Gerrad Y. Visualising junk: big data visualisations and the need for feminist data studies. Journal of Communication Inquiry. submitted.

 

@rosemarylhill

This entry was posted in Blog.

© Copyright Leeds 2017