Personalized Social Systems for Local Communities

About the project

The project explored the use of personalization and mobile computing to increase user engagement in location-bound social systems. The project was originally funded by a Research Award from Google to Peter Brusilovsky, however it continued for many years following the original round of funding. In the course of the project we developed, extended, and explored several location-bounded systems such as Comet, Eventur, and CourseAgent. These systems were actively used for many years in local Pittsburgh community. We developed a range of new recommendation technologies for these systems and explored new approaches to increase user engagement, which is the source of knowledge in local communities

Motivation

Social computing systems can serve as a tool for revitalizing and strengthening local communities - in particular by supporting online community information commons for enriching and actively disseminating information about local events, activities, and organizations. While existing social linking systems can allow people to follow or meet each other, they tend to do so in decontextualized virtual spaces and as a result are seen by many as a threat to local communities. In contrast, a social information system that provides a community information commons for collecting, organizing and exploring information about local events and organizations, has the potential to increase individuals’ awareness and connection with their community. However, creating an effective community information commons is challenging.

First, information about local events, activities and organizations is often fragmented, incomplete, and difficult to find. While existing channels work well for organizations able to devote resources to formal marketing and advertising, a sustainable community information commons must allow a more diverse set of community members to participate in collection and organization information in a decentralized fashion. Social computing technologies and techniques have the potential to provide a platform through which the small efforts of individuals, both within a community and beyond, can be mobilized to increase the quality of information available about local events, activities, and organizations. However, in spite of their potential, engaging users in social computing systems can be challenging. Users are eager to contribute for their own good (i.e., list an item to sell on Craigslist), but very hesitant to do it for a community good. Local community information systems must pursue different engagement strategies if they are to provide a useful, organized source of information about local events and activities.

At the same time, any system that is even moderately successful at aggregating information about community events has to be personalized. No one is interested all events and organizations. Even in a small community, the ability to focus in items of interest within a larger information stream is necessary. A viable community information commons must support personalized information access, enabling individuals to increase the likelihood of seeing personally interesting materials and decrease the distraction of irrelevant events.

The goal of this project is to investigate the nature of community information commons and develop social computing technologies and techniques to increase the impact and viability of these critical socio-technical systems.

Systems

• Eventur (formerly PittCult) more
• CoMeT
• CourseAgent

Team

• PI: Peter Brusilovsky:

• Rosta Farzan: Leading research on user engagement, the main developer of CourseAgent 2 system

• Brian Butler: Leading research the work on urban communities

• Chirayu Wongchokprasitti: Research on recommendations, main developer of the CoMeT system

• Danielle Lee: Research on recommender approached based on social links, main developer of Eventur system

• Claudia Lopez: Researcher on incentive-based design and user enagement

Publications

• Brusilovsky, P., Parra, D., Sahebi, S., and Wongchokprasitti, C. (2010) Collaborative Information Finding in Smaller Communities: The Case of Research Talks. In: Proceedings of 6th International ICST Conference on Collaborative Computing: Networking, Applications and Worksharing, Chicago, Illinois, USA October 9-12, 2010.
• Sahebi, S., Wongchokprasitti, C., and Brusilovsky, P. (2010) Recommending research colloquia: a study of several sources for user profiling. In: Proceedings of the 1st International Workshop on Information Heterogeneity and Fusion in Recommender Systems (HetRec 2010) at the 2010 ACM conference on Recommender systems, RecSys '10, Barcelona, Spain, September 26–30, 2010, ACM, pp. 32-38
• López, C. and Brusilovsky, P. (2011) Adapting Engagement e-mails to Users' Characteristics. In: Proceedings of Workshop on User Models for Motivational Systems: The affective and the rational routes to persuasion (UMMS 2011) at UMAP 2011, Girona, Spain, 11 July 2011, pp. 71-83.
• López, C. and Brusilovsky, P. (2012) Towards Adaptive Recruitment and Engagement Mechanisms in Social Systems. In: L. Ardissono and T. Kuflik (eds.): Advances in User Modeling. Lecture Notes in Computer Science, Vol. 7138, Springer Berlin / Heidelberg, pp. 382-396.
• López, C. and Brusilovsky, P. (2012) Designing Adaptive Engagement Approaches for Audience-bounded Online Communities. In: Proceedings of Workshop on Design, Influence and Social Technologies (DIST) at CSCW 2012, Seattle, USA, February 11-12, 2012.
• López, C., Farzan, R., and Brusilovsky, P. (2012) Personalized incremental users' engagement: driving contributions one step forward. In: Proceedings of the 17th ACM international conference on Supporting group work, Group’12, ACM, pp. 189–198.
• López, C., Farzan, R., and Brusilovsky, P. (2013) What Influences the Decision to Participate in Audience-bounded Online Communities. In: Proceedings of iConference’13, Fort Worth, USA.
• Nov, O., Arazy, O., López, C., and Brusilovsky, P. (2013) Exploring Personality-Targeted UI Design in Online Social Participation Systems. In: Proceedings of the 2013 Annual Conference on Human Factors in Computing Systems, CHI'2013, Paris, France, April 27-May 2, 2013, ACM, pp. 361-370.
• López, C., Butler, B., and Brusilovsky, P. (2014) Does Anything Ever Happen Around Here? Assessing the Online Information Landscape for Local Events. Journal of Urban Technology 21 (4), 95-123.
• Wongchokprasitti, C., Peltonen, J., Ruotsalo, T., Bandyopadhyay, P., Jacucci, G., and Brusilovsky, P. (2015) User Model in a Box: Cross-System User Model Transfer for Resolving Cold Start Problems. In: F. Ricci, K. Bontcheva, O. Conlan and S. Lawless (eds.) Proceedings of 23nd Conference on User Modeling, Adaptation and Personalization (UMAP 2015), Dublin, Ireland, , June 29 - July 3, 2015, Springer Verlag, pp. 289-301.