CIS Colloquium, Feb 11, 2015, 11:00AM - 12:00PM, SERC 306

CIS Colloquium, Feb 11, 2015, 11:00AM - 12:00PM, SERC 306

Sometimes a cigar is not just a cigar: Exploring the emergence and evolution of offline social processes on topic-specific online communities

Spyros Angelopoulos , University of Nottingham

This study explores the dynamic interplay between social temporal networks and human behavior on online communities, and more specifically it is focused on the possibilities that exist for crossing or even transcending the boundary between existence in the real and virtual worlds. It presents the findings of a longitudinal study of an online community of cigar smokers, tracing the interactions of its participants over a period of eighteen months. The online community was created for the needs of the study, and populated through an online invitation system to track the initial relations among participants. The explicit attention to network dynamics and the time dimension distinguishes this study from others on human behaviors on online communities: i) the study follows the day-to-day interactivity of the participants on the online community for eighteen months and conduct longitudinal social network analysis, and ii) approximates the death of non-reinforced or non-reciprocated connections. Studies to date effectively assume that once a connection between two individuals within the network is established, it persists forever. Whilst such a concept is not necessarily consistent with reality, the death of a connection is hard to identify in studies that are not longitudinal. This issue is addressed by tracking the persistence and death of connections among the interacting individuals. While on-going social interactions can produce observable peaks, by applying a lifespan filter, an approximation of the instantaneous network can be created. The instantaneous strength of a relationship is approximated by the average geometric rate of bilateral interactions within a lifespan window. Such an approximation allows the network of interactions to be recovered at arbitrary points in time by including only the ties with positive instantaneous strength. The lifespan window sets a relevancy horizon for identifying those events relevant to the current state of the network. The study identified six distinctive technology-driven social trends that affect the emergence and evolution of social networks among the participants. The distinctive patterns of interaction that persist over the course of the study are associated with a mix of behaviors that include play, trading and gifting, and entail the exchange or flow of informational and material objects. Moreover, the findings of the study shows that the offline interactions among the participants of the online community gave rise to a dense network of a homogenous population, with the properties of a scale free network, and of a small world with three degrees of separation. The interactions among the participants were highly reciprocated and reinforced, contributing to the growth of the network over time, and the tendency of participants to connect with friends of friends is equally spread in the network and not affected by prominence. There is a positive and statistically significant rich-club effect in the network, showing that the prominent participants do not compete with each other for status, rather they tend to interact with each other. The growth of the network can be divided into two periods: an initial accelerated growth, and an equilibrium period of homogenization of the population. During the initial accelerated phase, the backbone of the network is established, which contributes to the stability in shortest-path based metrics over time, and the overall density of the evolving social network. The findings of the study suggest that we should treat with caution conclusions that relate differences in status to differences in network position. Although hubs in a network can control the diffusion of recourses across the community, the unstable nature of such positions suggests that the participants do not hold them indefinitely, and these positions offer only a temporary advantage to those who posses them. Furthermore, it is unclear to what extent individuals can manipulate their positions in a network, even if that is their intention. The diversity of activities across the community, and the flow of information, cigars, and money demonstrate the potential for complex, multi-faceted socio-economic spaces that bridge the divide between virtual and embodied space, informational and material goods, and social and economic transactions.

Spyros Angelopoulos is currently a Research Fellow at the Horizon Digital Economy Research Institute of the University of Nottingham, in the UK. His interdisciplinary research lies on the intersection of Social and Computer Science, transcending traditional research boundaries. His current research is focused on the emergence and evolution of complex systems, and his research interests include Social Networks, Big Data Analytics, and Cloud Computing. He holds a PhD in Information Systems and Management from Warwick Business School, in the UK, a BA/MEng in Production Engineering and Management, as well as an MSc in Management Engineering from the Technical University of Crete, in Greece.

© 2001-2013 Center for Data Analytics and Biomedical Informatics, Temple University