[Research notes] Visualising PolySocial Reality

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.

PoSR network

Figure 1. An ‘exploded view’ of a fragment of a PoSR network. Each layer represents a different social network from the same set of individuals, each based on a different communication channel.

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)


Fig. 2. Multigraph integrating two networks.

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.


Fig. 3. Metagraph.

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.


5 thoughts on “[Research notes] Visualising PolySocial Reality

  1. I have to admit that I have problem in assessing the value of this contribution. The PolySocial Reality approach is new to me and drawing only on the present contribution, it is difficult for me to understand its interest and its specific focus. The only thing that I can say is that the contribution is very abstract and that one or more practical examples should be given to make clear how the PolySocial Reality approach could be operationalized. The link with the idea of ‘Just in Time Sociology’ should also be explained.

  2. Thanks for this comment. This is a new research area, and we are aiming to present a general outline of the theory of PolySocial Reality (PoSR). The next step for us is to test the theory with some large, online, real-time datasets, as well as conducting some qualitative (ethnographic) case studies in the real world. If selected we expect to present some practical examples of the theory at the workshop.

    One aspect we have not mentioned in our submission here (for brevity and simplicity), is one particular methodology we will utilise to investigate PoSR: user-generated personal and collaborative trails – through both digital and physical information spaces. There is some previous work on this in online e-learning systems, as well as in museums with mobile devices. We think this can be useful for exploring the theory of PoSR; and in turn, PoSR can inform trails which are multi-threaded instead of purely linear.

    I hope that provides a look at practical examples, as well as how it links to Just-in-time sociology, which tries “to understand social phenomena as they unfold, mining their digital traces.”

  3. Let me start by thanking the fine folks @ Anthropunk for this contribution. I share some of Tommaso’s perplexities. On the other hand, I tend to apprehend the notion of “polysocial reality” as a useful – albeit rather embryonic – tool to address what social scientists have described as “multiplexity”. In a 2005 contribution, Caroline Haythornthwaite (“Social networks and Internet connectivity effects”. Information, Communication & Society, 8(2): 125-147) defines relational multiplexity as the fact of belonging to different social networks, while media multiplexity is the overlapping of different analogically and digitally-mediated social networks. Prominently, social media are used to negotiate tie formation, strength and structure in both online and offline networks of contacts. This, of course, resonates with the idea of polysocial reality, yet it represents the theoretical framing for empirical data collected by social scientists via surveys (Haythornthwaite was working on distance learning, too). I’m afraid the present research note goes the other way around: authors have a theoretical framework, and want data to validate it (or just to tinker around with the idea). This theory-driven approach is consistent with some of the authors’ “maker” posture. But, in order to make it fit with the JITSO research orientations, it would be important for the authors themselves to identify one socially-relevant, politically significant, currently happening event allowing the collection of data. So, instead of apparently harvesting random online data via Twitter, Flickr, Last.fm, Foursquare and Ning APIs, I recommend to establish clearly a research question that allows to target a data set, focus on specific ties, and check for relevant hypotheses. In short, I encourage the authors to engage in the pursuit of authentic “citizen science” by coming up with a substantive social question, delimited in time and space, which will eventually shape their data collection and provide matter for their visualizations.

  4. Fair enough comment that we didn’t give specific examples of what we might do with PoSR in the text. Current applications are in UI design and applied design contexts which the authors prefer not to discuss prior to separate publications appearing. PoSR is a important element in two projects in Anthropology, one which begins in February that investigates the role of texting networks on pastoralists’ activity in Tajikistan, and another in preparation examining how coping knowledge is being formed, adapted and transmitted in Greece under Austerity.

    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 mobilise the unshared information of others in the network. In PoSR we are concerned with information access and flow across partially overlapping social networks based on the same or different relationships. We are, in some cases, interested in how a composite social network projects to a single individual, their capacity to assess information they access through their projected network, and their understanding of the network. However, more broadly (and often), we are interested in the composite properties of dynamically interacting (synchronously and asynchronously) multilayered networks that unfold over time. Hence our interest in both multigraphs and metagraphs and combining these as a means of representing fragments of a PoSR network.

    What is referred to as media multiplexity is also related to our interests, and is subsumed within PoSR, which includes lots of other sources of information. We also aren’t concerned with only internet or mobile communications. The research note describes research to operationalise aspects relating to PoSR, in particular to evaluate hypothesised properties of local regions of PoSR space. PoSR while easy to describe, is impossible to populate. PolySocial Trails are simply bits of data about relationships and traversals by single and multiple individuals and sources that we can observe in an effort to identify and evaluate higher order emergent properties of PoSR. PolySocial Trails do not have to result from passive observation, but can be the record of ‘experiments’ we pose to our consultants while they are traversing their transactions. Among the applications of this (unmentioned in the note) are the development of new statistical methods where sampling is dynamic rather than predetermined, and where the ‘expected’ distribution to measure agains is constantly changing, where this change itself can be recorded, observed and analysed rather than just the original data points, and where we can represent the network in terms of change differentials rather than simple transactions.

    The mission statement of the issue,”JITSO 2012 will gather the most significant international researchers that try to understand social phenomena as they unfold, mining their digital traces. ” suggested to us that our piece was relevant because given that PoSR is dynamic and the result of constantly unfolding relationships between individuals resulting in a stream of data that must be analysed as it is created, and where we can assess the impact of contingencies and interventions.


    Michael D. Fischer

  5. Pingback: [Conference] « Just-in-time sociology  a new field for digital humanities? | «Pegasus Data Project

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s