Data Power Conference
Panel Session 3a): Visualising Data (Chair: Richard Rogers)
What Can a Visualisation Do? Power and the Visual Representation of Data
Helen Kennedy, University of Sheffield; Rosemary Lucy Hill, University of Leeds; William Allen, University of Oxford, and Giorgia Aiello, University of Leeds
The main way that people get access to increasingly ubiquitous data is through visualisations – ‘data are mobilized graphically’ (Gitelman and Jackson 2012). Some writers claim that we are witnessing a ‘visualization of culture’ (Beer and Burrows 2013), others that visualizations can promote greater understanding of data through data transparency (Zambrano and Engelhardt 2008). It is important, then, to trace how visualisations come into being, the resources on which visualisers draw to produce visualisations, and the ways in which visualisations are imbued with scientific objectivity and transparency. On the one hand, turning data into a visualisation is not an automated process. Rather visualisation is ‘a purposeful act’, the result of numerous decisions, which often result in a visualisation that ‘pretends to be coherent and tidy’ (Ruppert 2014). Latour (1986/2008) laments this, asking:‘where are the visualisation tools that allow the contradictory and controversial nature of matters of concern to be represented?’ We might also ask the same question of visualisations. But on the other hand, visualization practitioners believe they can ‘do good with data’ (Periscopic, nd) and they devise their own professional and ethical codes: they want to do their work responsibly, be true to their data, reveal their sources, interrogate incomplete datasets, and they lament the ways in which intermediaries influence the visualization production process. So how does what visualisers say about their practice square with concerns about ‘emerging forms of rationality’ (Tkacz 2014) around data, numbers and their visual representation?Where does power lie, and how does power operate, in and through data visualisations?
Emotional Data Visualisations in Public Space: A Critical Overview
Christopher Wood, Queen Mary, University of London
Data collection is becoming an increasingly common part of our everyday lives, whether it takes place with our without our explicit awareness. Emotional data holds a particularly interesting place in this process. Although gradations apply across cultures, our emotional life is traditionally internal, something which is experienced most intensely in a personal and subjective way. The collection of emotional data may challenge this subjectivity, especially when it is mapped and exhibited publicly as part of a commissioned media architecture installation. This paper offers a critical overview of this process using case studies where emotional and sentiment data is aggregated and exhibited in public space.
Numerous examples exist of emotional data presented as map visualisations. However, the dissemination of emotional data as objects or interventions in public space is less common. Two case studies are examined in detail. ‘Energy of the Nation’ (commissioned by EDF Energy) utilised a battery of lights placed on the London Eye during Summer 2012. The colour of the lights were defined by UK twitter sentiment towards the Olympics. ‘D-Tower’ (2005) is a public structure commissioned by the city of Doetinchem, Netherlands which changes colour according to online questionnaire responses from the townspeople.
Following Paul Dourish and Malcolm McCullough, technical systems are understood as being given meaning by the economic and political contexts of their commissioning, design and presentation. This paper explores the significance of media art technology as a destination for emotional and sentiment data. This is considered alongside an analysis of the economic, political and social contexts that define the data collection and exhibition environment.
Clickivism and the Quantification of Participation: Studying Anti-Nuclear Activists on Facebook with Quanti-Quali Data Visualisations
Dave Moats, Goldsmiths College
This paper reflects on the consequences of social media data production on social life as well as on the social life of methods. Social media is thought to open up new spaces of resistance for activists, but these online interventions are commonly dismissed as “clicktivism”. In this paper I will argue that activist participation in social media should not be understood merely as an impoverished way of organising offline protests but on its own terms, as a means of challenging and re-framing mainstream media messages. However, social media tends to frame participation in quantitative terms (likes, votes, ranked comments), creating a game in which mainstream media and corporate PR are always better resourced, through “astro-turfing” and sponsored posts.
Social media platforms also encourage the quantification of social science methods through so-called big data techniques which largely ignore the symbolic, affective, qualitative dimensions of these interventions. I will propose that we need new quanti-qualitative (Latour and Venturini 2010) methods which simultaneously grasp quantifiable traces and visual textual data to better understand the power relations scripted into these platforms. I will use an experimental data visualization to explore these ideas through a comparison of anti-nuclear activists and nuclear PR Facebook pages. I find that resistance to dominant media actors is possible by hi-jacking their media streams, so long as we do not judge the success of these interventions in terms of 'likes' alone. I will also make some tentative reflections on the ethics of social media data collection between quant and qual.
Data Stories: Visualising Sensitive Subjects
Anna Feigenbaum, Dan Jackson and Einar Thorsen, Bournemouth University
The move toward Open Data brings with it opportunities for information re-use, increases transparency, and encourages civic participation in data analysis and communication (Graves and Hendler 2014). But while many datasets and digital archives grow bigger and more open, information on sensitive issues and vulnerable populations is far from ‘infinite.’ Working at these interstices there are often no straightforward data source, documents are scattered across agencies and organisations. Moreover, this kind of data is often kept hidden, deemed too ‘confidential’ to be made open. Such ‘uneven transparency’ raises important questions about the duty to document (Larsen 2014), particularly in regard to issues of security where obtaining information on vulnerable populations (prisoners, detainees, those living in conflict zones) becomes difficult.
Drawing from our Bournemouth University based Datalabs Project, this paper explores challenges that arise when working with data that is hidden, sensitive or obscured. Our Datalab project partners are organisations that investigate military and policing technologies, human rights violations and corporations with damaging ecological practices. Working on – and with – such sensitive subjects, means that storytelling with data comes with increased risks. In this paper we draw from our collaborative practices of co-creating data visualisations from these ‘difficult datasets’ to examine storytelling and visualisation techniques that can enhance the impact of data communications. Alongside this we reflect on the ethical responsibility researchers’ carry to consider the agency of vulnerable populations and the specific socio-economic and political contexts in which their subjectivities are articulated when we create narratives out of numbers (Aaron).