Sally Applin, Michael Fischer(Centre for Social Anthropology and Computing, University of Kent, UK)
Kevin Walker (Information Experience Design Programme, Royal College of Art, UK)
Physiologically, all humans sense, move and communicate in similar ways. But different cultures and sub-cultures interpret signals and interact with each other differently. Thus even without digital media, we inhabit what are often experienced as different, or even multiple, realities (see Spiro, 1986 for discussion). Within digital media this distinction has been used to differentiate the virtual from real-world experience (cf. Turkle, 2011). However, the Internet also spreads awareness of individual cultural practices, while simultaneously creating and facilitating new behaviours that cut across cultures. When we remotely access data and interact with others, while simultaneously moving and interacting in the physical world, our experience of reality become multi-threaded, or ‘poly-social.’ As a result, our very conceptions of time and space have become personalised, with the often asynchronous nature of online communication, and the need to be ‘somewhere’ in a specific time and place to communicate is diminishing.
New analytics and visualisation tools can enable us to depict and extract new meaning from PolySocial Reality, and our findings in turn can contribute to the development of new and existing tools and technologies, specifically through the integration of real-time datastreams and the interoperability between physical and digital contexts. Thus, the tools can inform social theory in an iterative fashion. Our research is framed as a form of citizen social science.
PolySocial Reality as a conceptual framework
As we incorporate digital social relationships, as well as personally-accessed content, into our lives through mobile and embedded technologies, social scientists have been developing theories and methods for studying online social and cultural practices. ‘PolySocial Reality’ (PoSR) is a conceptual framework proposed by Applin and Fischer (2012; 2011) to describe the emergent network comprised by the union of all individual networks such that patterns in the overall graph representing these can be identified, node-centric projections examined, and sub-graphs compared. Simply put, it helps identify the extent and impact of shared and unshared experience when people are interacting in social networks (both analogue and digital).
We are interested in the role and impact of unshared knowledge of cooperative tasks, given that people are engaged concurrently in many networks without other members of each network necessarily being aware of these interactions, and more important, the local mix of information that arises from these intersubjectively ‘masked’ connections. In particular we are interested in the use of mobile Internet to integrate social networking into daily life, which potentially compounds this PosR context even more, particularly in local analogue social cooperation where different individuals are interacting concurrently with a range of online social networks.
These issues are directly relevant to how mobile app developers conceptualise the network ecosystem, and we aim to provide evidence for a hypothesis that the degree of shared experience in the context of online linked individuals would increase where these connections were previously sparse, and decline where previously dense.
As we have developed the framework to date, it is apparent that it has potential applications for better understanding off-line networks as well. The practical applications of understanding the global and individual impact of the complex system of interactions represented by PoSR are potentially great, both with respect to improving users quality of experience, and the capacity of people to collaborate and collectively contribute to meeting the challenges that arise from social media.
Although PoSR has been utilised as a conceptual framework for understanding the contexts, connections and limitations of common information between individuals, to go much further in understanding the complex relationships people are embedded within requires data. There are a raft of problems in acquiring, analysing and representing this data. By looking at local graphs representing the activity of individuals who are partly connected through single and multiple social networks, we can identify some properties of local projections of PoSR, to inform hypotheses about individual impacts and the aggregate of these across PoSR more generally. However, empirically populating with data a complete PoSR structure, or even a reasonable sample, is for all practical purposes impossible without large datasets on the one hand, and tools for analysis and visualisation on the other. Thus our next step is to utilise publicly available online datasets and analytics tools to test and develop the PoSR framework.
Researching PolySocial Reality
We are initially addressing three basic questions:
- Which online data sources and behaviours best inform the theory of PolySocial Reality?
- How do these data sources map onto PolySocial networks?
- How are pathways through complex interactions—PolySocial trails—best visualised for data analysis and understanding?
In our research we harness data from different sources to inform our understanding of poly-social networks, including location and other contextual cues, search and trend data, and content that people access and share. To address our first Question, about the types of data to test our hypotheses about PoSR, we are designing a series of brief case studies to investigate different types of data to inform PoSR. These will be comprised of, first, an analysis of quantitative online data, including location, search content and trend data, which are secondarily and individually analysed in subsequent self-contained scenarios with a voluntary sample of users, in order to collect qualitative data. Data sources initially include sites with interactional data APIs such as Twitter, Flickr, Last.fm, Foursquare and Ning, and later adding a bespoke project social portal that aggregates participant activity on agreed channels for our voluntary participants.
We view the qualitative data collection as a form of participatory or ‘citizen science.’ which has gained prominence in recent years as a means to, for example, classify galaxies, fold proteins, find new planets, or identify objects or historical sources; by using human capabilities for pattern recognition it complements computational approaches to analysing quantitative data. Previous research undertaken by members of our team has investigated the value and dynamics of participatory e-science (See Smith et al, 2009). Our current research extends this work, and is the first we are aware of in the area of citizen social science: utilising voluntary users to contribute qualitative data to inform social theory. For us, such data is intended to illuminate particular micro-level behaviours; at a certain scale it will enable us to extrapolate macro-level social and cultural practices. Together these are intended to build up a rich picture of PoSR.
Data Representation, Visualisation and Analysis
To empirically investigate PoSR we need a means of sampling user activity and a means of flexibly representing the relationships we sample. Specific sampling methods, and the means of assessing samples, is critical. At the base, we conceptualise the structure of different social networks as dynamic graphs. Since we cannot enumerate all the potential people interacting in a given network at any one time, standard ‘crawling’ or ‘snowball’ sampling will not work, as this is known to result in biased samples on incomplete graphs (Lee et. al. 2006; Mislove et. al. 2007). Systematic random walks can produce a reasonable sample (Rasti et. al. 2008), but degrades in representativeness when the underlying graph is not highly connected. Multigraphs support one of the aspects of PoSR: the multiplicity of different relations that may underly multiple intersecting social networks. Gjoka, et al (2011) propose a promising approach to this problem based on multigraphs: graphs whose vertices may be connected by more than one edge. Because of the multiplicity of edges, each type representing a different context for social relations, a projected multigraph will be more likely to have a higher degree of connectivity. (See Fig. 2)
Combining multigraphs with metagraphs (Basu and Blanning, 1992) appears a reasonable initial mathematical representation for an exploration of PoSR through the analysis of trails and aggregated trails. A metagraph is a graph where vertices are sets, and edges are connections between sets (see Fig. 2). A multigraph that is also a metagraph permits us, at least, to represent the data in a form that is interoperative and can be converted into different forms such as matrices, XML or relational data suitable for online analytic tools for which a range of algorithms for methods of analysis have been established.
Because our meta-multigraph can be easily translated into more conventional projections of the data a range of existing open source tools for visualisation and analysis can be employed, such as Cytoscape (Smoot et. al 2011), GraphViz (Ellson et. al. 2003), Jung (O’Madadhain 2003) and R (R Development Core Team 2008).
One area of specialism in our team is data visualisation and simplification, with a specific focus on scientific data. We are additionally undertaking a process of rapid prototyping of concepts and tools in iterative fashion, as a form of design research. Existing analytics and visualisation tools, as well as the models described above, serve as a starting point for this design research.
We are aiming at both social scientists and the developer community with our tools and our framework, methods for using existing online data analysis and visualisation tools, and new tools and visualisation developed as a result of merging PoSR, trails, multigraphs and metagraphs. Such findings and tools will be made available online at PolySocial.netand open source repositories.
Applin, S.A. and Fischer, M.D. (2012). PolySocial Reality: Prospects for Extending User Capabilities Beyond Mixed, Dual and Blended Reality. Workshop on Location-Based Services in Smart Environments (LAMDa’12), in Proceedings of the 17th International conference on Intelligent user interfaces (Lisbon, Portugal, February 14-17, 2012) IUI ’12. ACM, New York, NY, 393-396.
Applin, S.A. and Fischer, M.D. (2011). A Cultural Perspective on Mixed, Dual and Blended Reality. Workshop on Location-Based Services in Smart Environments (LAMDa’11), in Proceedings of the 16th international conference on Intelligent user interfaces (Palo Alto, CA, February 13-16, 2011) IUI ’11. ACM, New York, NY, 477-478.
Basu and Blanning (1992) Enterprise Modeling Using Metagraphs. In T. Jelassi, M. R. Klein, and W. M. Mayon-White (Eds.), Decision Support Systems: Experiences and Expectations. Amsterdam: North.
Borwn, M. F. (2008) Cultural Relativism 2.0. Current Anthropology 49:3, pp. 363-383.
Ellson, J., E. R. Gansner , E. Koutsofios , S. C. North , G. Woodhull (2003). Graphviz and dynagraph – static and dynamic graph drawing tools. URL www.graphviz.org, accessed 8-10-2012.
Gjoka, M., C. T. Butts, M. Kurant, A. Markopoulou (2011). Multigraph Sampling of Online Social Networks. arXiv:1008.2565v2 [cs.NI]. http://arxiv.org/abs/1008.2565v2 (accessed 20 Sept 2012).
Lee, S. H., P.-J. Kim, and H. Jeong (2006) Statistical properties of sampled networks. Physical Review E, vol. 73, p. 16102.
Mislove, A., M. Marcon, K. Gummadi, P. Druschel, and B. Bhattacharjee (2007) Measurement and analysis of online social networks. In Proc. 7th ACM SIGCOMM Conf. on Internet measurement, San Diego, CA, pp. 29–42.
O’Madadhain, J., D. Fisher, S. White, and Y. Boey (2003). The JUNG (Java Universal Network/Graph) Framework. Technical Report UCI-ICS 03-17. School of Information and Computer Science University of California, Irvine. URL http://www.datalab.uci.edu/papers/JUNG_tech_report.html, accessed 8-10-2012.
R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org, accessed 8-10-2012.
Rasti, A. H., M. Torkjazi, R. Rejaie, and D. Stutzbach (2008) Evaluating Sampling Techniques for Large Dynamic Graphs. Univ. Oregon, Tech. Rep. CIS-TR-08-01.
Smith, H., J. Underwood, G. Fitzpatrick, K. Walker, J. Good, R. Luckin,D. Rowland and S. Benford (2009) Sustainability requirements for online science communities and resources. CAL’09 conference , 23-25 Mar 2009. http://www.esrc.ac.uk/…/4b1d37ee-7ead-432a-b636-c9201da52bceShare (accessed 30 Sept 2012).
Spiro, M. E. (1986). Cultural Relativism and the Future of Anthropology. Cultural Anthropology 1:3, pp 259–286.
Smoot, M., K. Ono, J. Ruscheinski, P. Wang, T. Ideker (2011), Cytoscape 2.8: new features for data integration and network visualization Bioinformatics. 27(3): 431–432.
Turkle, S. (2011). Alone together: Why we expect more from technology and less from each other. New York: Basic Books.