Sally Applin, Michael Fischer(Centre for Social Anthropology and Computing, University of Kent, UK)
Kevin Walker (Information Experience Design Programme, Royal College of Art, London)
Physiologically, all humans sense, move and communicate in similar ways and have the same sensors, but different experiences (Applin and Fischer, 2011). Furthermore, as humans socialise within different cultures, they interpret signals and interact with each other in vastly different ways. This heterogeneity of human culture happens in both analog and digital communications. Turkle (2011) has used this distinction to differentiate the virtual from real-world experience; however there is some debate as to whether or not a clear distinction of only two worlds can be drawn in such a dualist manner (e.g., Jurgenson, 2011). The Internet as a human construction has enabled new capabilities for humans (Applin and Fischer, 2011), who are now able to spread 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 within our immediate locale, our individual experiences become multi-threaded, or ‘poly-social.’ As a result, our very conceptions of time and space have become personalised and mostly asynchronous, with the mobile phone often acting as a tool for applying a ‘just-in-time’ model for social planning.
PolySocial Reality (PoSR) describes the multiple, sometimes overlapping, network transaction spaces that people traverse synchronously and asynchronously with others to maintain and use social relationships, and has been developed as a theoretical framework for the global interaction context within which people experience the mobile social web and other forms of communication and social interactions, whether co-located or mediated by technology. We expect new analytics and visualisation tools to 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 as an iterative, ongoing, process. Our data collection methods include a novel form of ‘citizen social science.’
PolySocial Reality as a Theoretical Framework
PoSR is a theory proposed by Applin and Fischer (2012; 2011) that defines relations across the aggregate of all the experienced locations and communications of and between all individual people in multiple networks and/or locales at the same or different times. A simple example of a PoSR context occurs when two people are walking together down a street, while simultaneously texting or communicating through both digital and analog channels that are partially or wholly interleaved, or that replace face-to-face interactions with interactions in a single, dual or multiplexed fashion. PoSR is based upon the core concept that dynamic relational structures emerge from the aggregate of multiplexed asynchronous or synchronous interactions of all individuals within the domain of networked, non-networked, and/or local experiences. We intend to use PoSR to describe and analyse principles underlying instantiations of the emergent ‘network’ comprised by the union of all individual networks, such that patterns in an overall graph representing these can be identified, node-centric projections examined, and sub-graphs compared. Simply put, PoSR helps identify the extent and impact of shared and unshared experience when people are interacting in social networks (both analog and digital).
Primarily we are concerned with whether or not these asynchronous and synchronous multiplexed and/or individual messages are received and acted upon in a way that facilitates human cooperation. If too many messages become overwhelming, meaning can be lost, and lost meaning can manifest, among other things, as misunderstandings that have consequences going well beyond whether or not someone got the message to pick up milk at the store. In the United States, the transportation industry has been transformed by new regulations, as automobile drivers, train engineers, barge captains and airline pilots have all missed messages that resulted in complications from the fatal, as in a barge disaster in Washington D.C., to humorous, as in the case of airline pilots missing their airport by 300 miles, while being distracted by an iPad.
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 and opportunities arising from social media.
To go further in understanding the complex relationships PoSR represents requires data. There are a raft of problems in acquiring, analysing and representing this data. However, by looking at local graphs representing the activity of interacting 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. Empirically populating with data a complete PoSR structure is not possible. However, a sample can be drawn from sufficiently large datasets, and examined with tools for analysis and visualisation. Thus our next step is to use publicly available online datasets and analytics tools to test and further develop the PoSR framework.
Researching PolySocial Reality
We are initially addressing three basic questions:
- Which data sources and behaviours best inform the theory of PolySocial Reality?
- How do these data sources reflect PolySocial networks?
- How could pathways through complex interactions—PolySocial trails—be utilised for data analysis and understanding of formative principles underlying PoSR?
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 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 public 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 phase 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 museum artifacts or historical sources; by using human capabilities for pattern recognition, this 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 flexibly representing relationships – both co-located and digitally networked. We conceptualise the structure of social networks as dynamic graphs. Simple social graphs have been used to study the structure of online social networks (e.g. Mislove et al, 2007), as depicted in Figure 1.
However, this representation says nothing about the form or content of these exchanges. When multiple channels or media through which people communicate (including face-to-face) are considered, this can result in a more complex or ‘multiplexed’ network of exchanges. Additional parameters include the strength and quality of ties, the types of exchanges, frequency and duration of contact, and degree of relatedness (See Haythornthwaite, 2005). Gjoka, et al (2011) address this through multigraph sampling. Multigraphs – graphs whose nodes may be connected by more than one line – support one of the aspects of PoSR: the multiplicity of different relations that may underly multiple intersecting social networks. Because of the multiplicity of connections, each type representing a different context for social relations, a projected multigraph will be more likely to have a higher degree of connectivity. For example, the simple scenario described above, of two co-located individuals, can be represented with multiple connections to each other; they may be connected as friends as well as being co-located in a particular event, as well as through a social networking service, which connects them as well to others (Figure 2).
Figure 2 shows how a simple instance of PoSR can get complex very quickly. However, combining multigraphs with metagraphs (Basu and Blanning, 1992) appears a reasonable initial mathematical representation for an exploration of PoSR. A metagraph is a graph where each node is a set, and edges thus connect sets (see Fig. 3). A multigraph that includes metagraphs 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.
Our proposed meta-multigraph can be easily translated into a wide range of more conventional projections of the data, and 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 et. al. 2003) and R (R Development Core Team 2008).
One way of exploring this data representation is through trails and aggregations of trails. The concept of trails through informational ecologies was initially proposed by Bush (1945) and developed by Peterson and Levene (2003) as a form of navigational or ‘ampliative’ learning. Originally conceptualised for representing individuals moving through an informational space, trail visualisation was developed by Schoonenboom (2007) as multigraphs as well as activity diagrams and interaction maps. Walker (2012) has investigated trails using mobile technologies, in which individuals are conceptualised as situated in overlapping personal, social and physical contexts (see Fig. 4)
Applying this to our original model, this enables us to regard two individuals as nodes in a multigraph, situated in particular contexts, with their communications mediated by available tools and resources. This approach thus opens the possibility to include contextual data such as location in our analysis (Fig. 5).
We aim to investigate trails – essentially linear paths through nonlinear graphs – through meta-multigraphs representing ecologies of networked individuals. PoSR is a model that includes multiplexity as a basic network property. We are looking, however, at the properties of subnetworks whose nodes have differential distributions of multiplexity structure. In particular, we are interested in the relative density and distribution of information between nodes in a PoSR fragment based on the extent of common shared nodes in a multigraph, and at the impact of this of mechanisms for mobilising the unshared information of others in the network.
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 posr.org and in 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.
Bush, V. (1945) As we may think. Atlantic Monthly, July 1945.
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).
Jurgenson, N. (2011). Digital Dualism vs Augmented Reality. The Society Pages. Cyborgology Blog. (February 24, 2011.) Retrieved December 2, 2012 from http://thesocietypages.org/cyborgology/2011/02/24/digital-dualism-versus-augmented-reality/
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.
Peterson, D. and Levene, M. (2003) Trail records and navigational learning. London Review of Education 1(3).
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.