Olivier Phillipe, University of Southampton
Introduction
The recent explosion of digital information and the consequent ‘data deluge’ [11] has had a profound effect on business practices, particularly marketing, as organizations are increasingly able to exploit administrative and transactional traces of customer behaviour in order to develop ways of understanding their markets and selling direct to users (e.g. in Amazon’s ‘if you liked this you might like …’ or Tescos bespoke marketing via Clubcard) [14]. Meanwhile, however, the recent rapid rise in use of social networking sites offers digital traces of a far broader nature - not limited to commercial/transactional interactions between a company and its customer but generating content about users themselves and their interaction with others – creating a ‘social stream’ of data. Furthermore, some of these data are freely available for analysis, most notably on Twitter.
In this paper, we explore some of the opportunities and challenges involved in working with the social stream. In distinction to most debate in this broad area we do not concentrate on the methodological issues involved [2, 8]. Rather, we argue that big data raises fundamental ontological, epistemological and theoretical issues for social research. For whilst there are surely challenges involved in harvesting and visualising the social stream [4] we must also consider the very nature of these data, what claims to knowledge might be made from these data, and what kinds of concepts and theories are appropriate to describe and analyse them. In sum, we will argue that it is important to grasp the distinctive nature of ‘big data’, the challenge that this poses to conventional social scientific categories and assumptions, and the consequences for theoretical development. In turn, we suggest that these considerations should inform the development of theoretically grounded methodologies and methods.
The Structure and Nature of the Social Stream: Challenge and Opportunities
The Social Stream may be characterised in terms of the 3Vs [13]: volume – the unprecedented quantity of social data; variety – the heterogeneous forms that these data take; and velocity – the constant renewal and flow of these data. Taken together, these characteristics comprise a data-complexity that poses major challenge for social scientific research, not least in how to apprehend and visualise the structure of these data.
At the same time, the Social Stream promises considerable opportunity to social scientific research: the social stream is naturally occurring data, spontaneously generated information about users and their interactions with others; offering data on actual behaviour, in real time, rather than behaviour reported in surveys or interviews; in a variety of formats, from 140 character tweets to online video performances; on a variety of subjects, from collapsing political dictatorships to celebrity gossip).
This data-complexity – both in terms of structure and content - demands a radical turn in methodology and theory. In order to visualise the structure of these data, and their dynamism, we must draw on computational techniques; and in order to analyse these data we must engage with theoretical perspectives which enable us to recognise and interrogate the fragmentary, fluid and dynamic nature of these data. Drawing together the methodological and the theoretical, the key issues are as follows: (1) Activity (2) Post-demographics (3) whole population.
1. Activity
Focusing on the activity as the relevant available information in Social Stream is only possible through a certain theoretical lens emphasis on the dynamic. Theses theoretical tools already exist to analyse the dynamism of social interactions instead of measuring their stability. Since the emergence of informational age [5] different theories mobilize concepts such fluidity [15] liquidity [1] or mobility [12]. More recently the actor-network theory expresses the importance of such perspective by stating that only activity and trace left by people can be investigate as a manifestation of social interaction [6]. But these theories pose some limits in order to fully understand activity such as context for actor-network theory [7] or the necessity to follow the travellers in mobile sociology [3]. This is why we need to develop the concept of post-demographics, initially coined by Rogers to urge sociologist to develops new methods to study phenomenons occurring on Social Network Sites [9].
2. Post demographics
People are posting about the preferences their tastes and their interests, they are building networks with people matter for them. This post-demographics information differs from our standard social science instrument, it is not driven by pre-defined categories – class, gender, race but is driven by the user activity itself. The categories we can built from the post-demographic data permits us to shift from a global perspective centred on demographic information toward a contextualised subjective perspective.
3. Whole population
However, in order to draw categories based on post-demographic data, and given the volume of data, we need to apply specific analysis in order to retrieve relevant information. For instance, doing a social network analysis or applying a cluster analysis is considering the data as a set and not aiming on any point a statistical representation of a larger population. The techniques employed to analyse them consider the set as a whole population, even if they are de facto a sample of a larger dataset. It is here precisely that the interaction between data complexity (computer science perspective) and social complexity (social science theoretical tools) gives us the opportunity to see over the demographic information.
This question of dynamic activity (velocity) contextualised with post-demographics data (variability) on a massive scale (volume) will only be answered by taking an interdisciplinary perspective. As well as the interviews techniques evolved consecutively to the theory in sociology [10], we need to understand how computer techniques will shape our further analysis in terms of limits and possibilities and how, in return, our theoretical tools will impact our understanding of data collected.
References
[1] Bauman, Z. 2000. Liquid modernity. Polity.
[2] Boyd, D. and Crawford, K. 2011. Six Provocations for Big Data. (2011).
[3] Büscher, M. and Urry, J. 2009. Mobile methods and the empirical. European Journal of Social Theory. 12, 1 (2009), 99–116.
[4] Callebaut, W. 2012. Scientific perspectivism: A philosopher of science’s response to the challenge of big data biology. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences. (2012).
[5] Castells, M. 1996. Rise of the network society: The information age: economy, society and culture. Blackwell Publishers, Inc.
[6] Latour, B. 2011. Networks, Societies, Spheres: Reflections of an Actor-Network Theorist. International Journal of Communication. 5, (2011), 796–810.
[7] Latour, B. 2005. Reassembling the social. Oxford University Press Oxford.
[8] Manovich, L. 2011. Trending: The Promises and the Challenges of Big Social Data. Debates in the Digital Humanities, ed MK Gold. The University of Minnesota Press, Minneapolis, MN.[15 July 2011]. (2011).
[9] Rogers, R. 2009. The end of the virtual: Digital methods. Amsterdam University Press.
[10] Savage, M. 2010. Identities and Social Change in Britain since 1940: the politics of method. Oxford University Press Oxford.
[11] Savage, M. and Burrows, R. 2007. The coming crisis of empirical sociology. Sociology. 41, 5 (2007), 885–903.
[12] Sheller, M. and Urry, J. 2006. The new mobilities paradigm. Environment and Planning-Part A. 38, 2 (2006), 207–226.
[13] Taming big data:http://www.ibm.com/developerworks/data/library/dmmag/DMMag_2011_Issue2/B.... Accessed: 2012-03-10.
[14] Thrift, N.J. 2005. Knowing capitalism. Sage Publications Ltd.
[15] Wittel, A. 2001. Toward a network sociality. Theory, Culture & Society. 18, 6 (2001), 51.
