Data is a hot topic. It is both something that will save us, and something we must be saved from. It is the next revolution, a deluge and a particularly potent ideology. On the one hand we have major digital institutions, fronted by social physicists and computer scientists, developing ever more sophisticated tools for storing, analysing and visualising data. On the other hand we have a growing number of critical scholars interrogating issues ranging from privacy concerns to the data hype’s ontological and epistemological baggage.
Recent seminars on ‘Data Practices’ – the theme of this years Design and Social Science seminar series co-organised by the Design and Sociology Dept at Goldsmiths – have tried to combine the affirmative and the critical approach. The projects presented so far both experiment with the practical possibilities that new data sources and devices open up but also reflect critically on the capacities of data in specific socio-material settings. In this way these projects try to move beyond an unproductive division of labour, according to which there are ‘naïve’ practitioners, on the one hand, and distanced critics, on the other.
Precisely because of this dual engagement with data, the seminars have been able to touch on a variety of different practical and theoretical questions. The classical epistemological questions have been supplemented (in some cases trumped) by questions regarding data ecologies and performativity. Moreover, when one thinks about data in its specific socio-material settings, one arguably foregrounds questions regarding the design of data. To paraphrase Kanye West, data needs to be architected.
In the introductory seminar titled “A thing to talk with”, Evelyn Ruppert and Noortje Marres presented data devices that intervene in the dominating ways of filtering and ordering information online. Ruppert is founding editor of the new journal Big Data & Society and her talk was about designing the journals new logo. The logo is going to be a network of keywords used by the journals authors to categorise their articles’ main themes, or as Ruppert characterized them, their ‘matters of concern’. In other words, the logo of the journal is meant to live stream the ‘matters of concern’ of the articles within the journal. The changing logo allows the journal both to have a specific profile, but also for this profile to change as the concerns of its contributions change. The nodes of the network are going to be the keywords and an edge is drawn between them if they are used to categorise the same article. The logo is then going to be a network of more or less connected matters of concern. Relevance is created on a relational principle. Representing the journal in this manner arguably goes against the common popularity indexes that either shows the most cited articles, the most read articles or the most used words.
Figure 1, Big Data and Society logo, made together with DensityDesign.
Marres’ thing to talk with was the Associational Profiler, developed by CSISP together with the Digital methods Initiative (University of Amsterdam). The profiler, like the logo, focuses on the connections between issue terms, but in this case these are derived from social media platforms like Twitter. Marres’ talk was focused on the technical and practical difficulties involved and the slightly chaotic process of developing the associational profiler. The profiler is not just an abstract machine that filters information according to certain principles, it is also a messy material and social arrangement. It is hardware filling up a university office, software and an interface written in certain programming languages and measures deriving from some methodological traditions rather than others. All parts of these arrangements are vulnerable and needs to be constantly worked on. Marres also talked about the difficulty of converting actor network theory’s principles of relevance into a functioning device. The Associational Profiler is in many ways an intervention and an experiment that might fail.
Figure 2, Associational Profiler’s visualization of hashtags co-occurring with #netfreedom on Twitter over time.
Where Marres and Ruppert focused on the building of information filters that attempt to intervene in dominant ways of ordering information online, Alex Wilkie and Jennifer Gabrys were more concerned with participatory data and design practices. Wilkie presented the Energy Babble, a machine that is designed for energy reduction communities by the Interaction Research Studio at Goldsmiths, which gathers voice and phone messages, energy usage updates from the UK grid, as well as scraped tweets concerning environmental issues and sustainable living. The Babble reads the data mashup out loud with a synthesized machinic voice, a re-mediation that complicates what communication around environmental issue is and can be.
Gabrys, together with other members of Citizen Sense Lab (Goldsmiths), presented their work on participatory environment sensing technologies. Particularly their experiments involving the Air Quality Egg, “a well-known DIY air quality sensor built up through a distributed maker community”, that promises easy involvement with environmental issues through data.
The marketing promise of easy involvement quickly proved false. For the device to produce more reliable data, Citizen Sense Lab had to take apart the Air Quality Egg, calibrate its sensors, play around with the code and re-engineer their own device. Thereby clarifying the hard work, the ‘not so easy involvement’, that goes into doing environmental research well. In this way th project intervened in participatory discourse and practice with their own critical engagement with DIY technologies and communities, touching upon a variety of critical issues around citizen science.
Experimenting with purpose
In the third seminar, working members of the Interaction Research Studio described the design processes of a number of their projects. A theme that cut across them was the emphasis on localities of data and how data can be embedded in concrete situations and objects.
One of the projects described was a portable object with a small screen that displays text messages, drawn from heterogeneous data sources, some generally accessible online, while others only granted access upon request (Wikipedia, Zoopla, the Census, NHS). What is shown on the screen of the device are snippets of data organized in respect to device’s geo-location (say the Burrough of Lambeth) and its relative speed of movement (i.e. is the user travelling on foot or by means of transport). Data, displayed with various speeds, is thus localized and contextualized in situ. Users of the device are presented with various disparate and randomly (in other respects than the data’s connection to location) sequenced pieces of data that inform them about the places they are moving through.
The randomly displayed data sequences can evoke interesting reflections in respect to the purposing of data. Data normally has a specific purpose. Governments are using data to manage populations, companies to target customers, and social scientists to produce knowledge about social life. What struck us about the IRS device-in-progress is that it enables a certain de-purposing of data, most apparent in the randomness of the information displayed. Here, data taken from heterogeneous datasets produced with determinate purposes, is not simply re-purposed and employed to do something else than in its original design. It re-mediation rather has the effect of detaching the information from any specific purpose. Data, in a sense, becomes a collection of incoherent objects that create a surreal situation, where components of different worlds are brought together in a given situation. “My neighbourhood has a relatively high unemployment, ”@JoeBadass tweeting about his lovelife,” and “a tendency to favour Pepsi Cola over Coca Cola.” This raises interesting questions as to what it means to locate something. What creates the distance between the situated user and information displayed?
The de-purposing of data can also be seen as a way of reflecting on data as an object instead of as a representation of an object elsewhere. That the device gives information that is not immediately relevant to you, has the effect of frontstaging data itself as a matter to engage with. As one of their sketches suggested, this technological arrangement enacts a data flâneur, who drifts through urban landscapes, aimlessly fishing for snippets of data attached to “current location.” Exploring data as cut-up poetry. Data, stripped from any preconceived purpose, in a way opens the ground for more creative re-purposing. What happens if we connect a database, through a designed device, to a certain location? What kinds of purposes can then be established? These are questions that can only be answered through practice. At times, experimental play with data may become a process of creating ‘data imaginaries’. A broader range of possibilities of data use can be activitated through material design practice.
By Goran Bečirevič and Hjalmar Carlsen