Social network analysis: Measuring tools, structures and dynamics

This special issue of Physica A presents contributions drawn from papers presented at a conference on Social Network Analysis and Complexity held from July 31 to August 2, 2006 at the Collegium Budapest in Budapest, Hungary. Co-chaired by Andras Lo+rincz of the Eotvos Lorand University and Nigel Gilbert of the University of Surrey, United Kingdom, the conference brought together scientists and scholars from mathematics, computer science, graph theory, statistical physics, network analysis, sociology, political science, technology and media. Papers were delivered on tools for social network analysis; abstract mathematical concepts, analogies, and algorithms; and methods for measuring the dynamics and structure of social network. Case studies were presented as well as papers on large-scale simulations of agent networks and artificial societies. A picture was presented of an area of research that is fertile with possibilities for productive inter-disciplinary collaboration, and where important results are beginning to emerge. In this issue, the contributions to the conference have been divided into three sections. First, there are papers that address the problem of characterising complex social networks. This is an area where it is particularly fruitful to bring together social scientists, who have long been collecting empirical data about actual networks, and statistical physicists, who have been wrestling with how to analyse network phenomena, but have generally been more concerned with the relatively orderly domain of particles than the apparently messy social world. Jost et al. develop an analysis of social inter-dependence that they suggest could be used to understand social differentiation, developing a theme that has been central to social science since Durkheim’s work in the late nineteenth century, but now with a complex twist. Chen et al. use ideas from percolation theory to develop a new measure of the fragmentation of a social network and apply it to data on links between workplaces. Vicsek uses tools from statistical mechanics to find overlapping communities in a social network. Shi et al. examine the conditions under which networks can transmit information efficiently and show that nontransitive links make very little contribution, testing their theory with data from real networks as well as with a random network. Ormerod shows, through an empirical example of access to financial services, that it is possible to make inferences about the structure of a network from rather limited information about the characteristics of a random sample of individuals. Finally, Roth stands back from these analyses and outlines an epistemological perspective on the application of statistical physics to the analysis of complex social networks. Second, there were contributions that were concerned with applying tools and methods to a variety of specific social networks. Pentland has collected data on the location and behaviour of groups of people, using