Social Roles in Online Networks
Why do users interact with others in an online social system? The sociological concept of a user’s role in a social network, which is a classification that takes into account social context and network position, is a key tool for unearthing the fundamental reasons we connect and interact with others. Role analysis forms the foundation to countless offline social network studies, yet the present methods for extracting social roles do not scale to the large-scale networks representing interactions in online social systems.
This talk will motivate the need for integrating sociological theories into modern computational methods and models of online social systems and discuss a new, data-driven approach for mining social roles in large-scale networks. The approach hinges on understanding the microscopic structures these networks are composed of: role triads, which are triples of users (conditioned by position) and the interactions among them. Each triad is born out of a unique social process, and hence, provides a basis for discovering the forces that underpin the network’s development. We find social roles in networks of interactions within Facebook, Stack Overflow, Wikipedia, and a private student social network from UC Irvine by clustering them in a space defined by number and types of role triads that compose their ego-networks. We explore the ego-network structures and role triads that best characterize various types of roles, and interpret them to find the social forces that drive user interactions across each network.
Derek Doran is a Ph.D. Candidate in the Department of Computer Science & Engineering at the University of Connecticut. His research interests are in large-scale and social network analytics, social media mining, and in the performance and reliability modeling of distributed systems. His current research focuses on understanding sociological patterns in large-scale online networks, social media mining with applications to the utility, transportation, and smart city industries, and understanding the performance implications of Web crawler traffic on the Internet. Derek is a National Science Foundation East-Asia and Pacific Summer Institutes Fellow, A Transportation Research Board Graduate Research Fellow, a twice Summer Alumni of Bell Labs in Murray Hill, NJ, a Summer Alumni of the Dependable Distributed Systems and Networks Lab at National Taiwan University, and holds an Interim Researcher position at ABB Corporate Research. His work has received Best Paper Awards at multiple recent conferences. He is a student member of ACM, IEEE, SIAM, and a young member of the TRB.