The adoption of computing and communication devices and networks has fundamentally changed the way in which people communicate with one another.
They use information services, blogs, wikis and other media to exchange information, opinions, expertise and knowledge. This has implications for organizations that provide these media and associated communication services, organizations that use these services, and for society. The social analytics center working in partnership with a consortium of firms has compiled a repository of unique data sets pertaining to user behavior in social information sharing contexts that will support fundamental and applied research.
Professor, Language Technologies Institute and Heinz College
Dean, Heinz College,
Professor of Management Science & Information Systems
The work of this center is organized around two thrusts:
- The convergence of high speed wireless networks and full featured hand held devices (e.g., smart phones) is enabling the integration of communication and associated information services into everyday business and recreational activities by individuals globally. New software platforms (e.g., Android) enable the creation of applications that work across devices and network service providers. Devices are personalized (e.g., caller ring back tones) and used to generate, consume, and share content (e.g., photographs, music, videos, micro-blogs). The patterns of sharing and communicating result in the formation of “mobile online communities” that is interesting from both a social and an economic perspective. We will address the following questions: How is the diffusion and adoption of new technology (e.g., new handsets) influenced by membership in these communities? Do members of the community exert social influence on one another and can this social influence be quantified? How can new information services that leverage social influence be developed? What are interesting societal and managerial applications of such technology? Working in close partnership with global mobile communication providers, the social analytics center will develop predictive models, analytic tools and techniques to answer these questions.
- The widespread use of blogs, wikis and question answering permit formal and informal organizations to tap into the “wisdom of the crowd”. Examples include societal communities such as Wikipedia and Yahoo Answers as well as their organizational cousins that fall under the broad category of knowledge management systems. Working with a consortium of global IT service providers, the center has compiled a unique data set of intra-organizational blogs, question answering forum behavior and document-authoring with the objective of characterizing expertise at the individual level. Using a collection of techniques for language analysis, text mining and information retrieval, the goal is to understand questions such as How opinion and knowledge diffuse through an organization? What is the relationship between physical network relationships and network relationships established through citation, replying and posting behaviors in blogs? and What are the alternative ways in which expertise can be characterized (e.g., bag of words) and how does the availability and access to such expertise determine outcomes, be they of organizations or of individuals?
Selected Research Findings:
The e?ect of the network structure on the dynamics of social and communication networks has been of interest in recent years. It has been observed that network properties such as neighborhood overlap, clustering coe?cient, etc. in?uence the tie strengths and link persistence between individuals. In this paper we study the communication records (both phonecall and SMS) of 2 million anonymized customers of a large mobile phone company with 50 million interactions over a period of 6 months. Our major contributions are the following: (a) we analyze several structural properties in these call/SMS networks and the correlations between them; (b) we formulate a learning problem to determine whether existing links between users will persist in the fu- ture. Experimental results show that our method performs better than existing rule based methods; and (c) we propose a change-point detection method in user behaviors using eigenvalue analysis of various behavioral features extracted over time. Our analysis shows that change-points detected by our method coincide with the social events and festivals in our data.