Do data visualizations, as they appear in public debate,work to clarify or conceal arguments? Do they lay themselves open to (in)72 torgeir uBerg nærlanDvalidation? And how are rhetorical devices used to convince? Such questionsare important to answer in order to attain a more critical understanding ofhow data visualizations contribute to public and political discourse—ormore generally; to manipulative or argumentative public spheres.

A very basic example is national weather forecasts. These visualizationsinvolve a simple form of recognition: the acknowledgement of peripheralparts of a country and the people who live in them as part of the country.Another example is political maps. Political maps involve the recognitionof sovereign territories as nation states, and the non-recognition of oth-ers—often those with unfulfilled claims to sovereignty. What is includedand what is not in data visualizations—who is given visibility through datavisualizations—thus emerges as a condition for recognition. This capacity ofdata visualizations to give visibility to groups or persons

For instance, a regular staple in these summaries has been the visualiza-tion feature called the ‘burning ember’ (see New York Times, 2009). In the formof coloured bar graphs, the ‘burning ember’ visualizes risks (the redder, themore critical) associated with different scenarios of increased global meantemperatures. As such, the ‘burning ember’ provides an example of howdata visualizations address ‘strong’ publics, whose discourse encompassesboth opinion formation and decision-making. It is thus also an example ofhow data visualizations may attain very concrete and manifest politicalsignificance as aids for political decisions. However, the inclusion of the‘burning ember’ has been criticized precisely for being too instructive.

An obvious capacity through which data visualization may enablecitizens is by providing them with information and with tools for makingsense of complicated political issues. It may enable citizens to take partin political will and opinion formation as well as to form informed partypreferences. Moreover, data visualizations may also provide valuable inputto the everyday and informal discussions among ‘ordinary people’, sittingat the core of deliberative models of democracy

Beer’s account, however, brings into attention an important distinctionbetween data visualizations as ideology and data visualizations as carriersof ideology. Beer is exemplifying the former. For Beer, who is inspired bythe discourse theory of Michel Foucault, data visualizations discursivelyconstitute the trust in numbers that is at the heart of neoliberal ideology.In its forms and modes of production, data visualization here embodies thevery logic of neoliberalism. In contrast, we can think of data visualizationas a tool for symbolic representation of issues of ideological significance.From this latter perspective, ideology is not contained in the form of datavisualizations themselves, but is a matter of what is represented and how

From this perspective, data visualizations are integralto the production of meanings, signs, and values in social life, and, ac-cording to Marxian thought, a vehicle for the legitimation of the ideas of aparticular group or class.

Weak publics, according to Fraser (1992, p. 134), are those publics ‘whosedeliberative practice consists exclusively of opinion formation and does notalso encompass decision making’. Weak publics would typically includenon-elite citizens and media audiences. Strong publics, by contrast, arethose ‘whose discourse encompasses both opinion formation and decisionmaking’ (p. 134) and may include politicians and bureaucrats. Consequently,the direct impact of data visualizations on political decisions will be strongerwhen they engage strong publics.

Crucially, thediscussion that takes place in the public sphere is, ideally, connected to actualdecision-making. The public sphere should be the mediating space betweenprivate persons, civil society, and political decision-makers (Habermas,2006). The core idea is that political decisions should be grounded in publicopinion, not only on electoral results and on the negotiations among eliteactors.From this perspective, data visualizations become significant as partof public discourse. Consider for instance how visualizations of carbonemissions data frequently are employed in public discussions about transportpolicy

Before we can outline these perspectives, it is necessary to make somegeneral demarcations about what we mean when we talk about politicalsignificance in the context of data visualizations. First, ‘politics’ may beunderstood in narrow terms, as the workings of political parties, processes,and institutions. Politics may also be understood in a wider sense, as thestruggle for power more broadly, as this struggle takes place both in theprivate as well as the cultural sphere, and by symbolic as well as materialmeans.