[Paper Review] Social Systems: Can we Do More Than Just Poke Friends?

Georgia Koutrika, Benjamin Bercovitz+5 more|ArXiv.org|Sep 9, 2009

Expert finding and Q&A systems12 references3 citations

TL;DR

This paper presents CourseRank, a social system for course evaluation and planning that goes beyond basic social interactions like 'poking' friends. By leveraging user-generated course reviews and collaborative filtering, it addresses novel research challenges in social data management, demonstrating that social systems can support complex, semantically rich applications with significant research potential beyond simple social features.

ABSTRACT

Social sites have become extremely popular among users but have they attracted equal attention from the research community? Are they good only for simple tasks, such as tagging and poking friends? Do they present any new or interesting research challenges? In this paper, we describe the insights we have obtained implementing CourseRank, a course evaluation and planning social system. We argue that more attention should be given to social sites like ours and that there are many challenges (though not the traditional DBMS ones) that should be addressed by our community.

Motivation & Objective

• To explore whether social systems can support more than trivial interactions like 'poking' friends.
• To identify and address novel research challenges in social data management beyond traditional database systems.
• To design and implement a real-world social system—CourseRank—that supports course evaluation and planning through user collaboration.
• To demonstrate that social platforms can enable complex, semantically meaningful interactions with significant research value.
• To advocate for greater research attention on social systems that support structured, domain-specific collaboration.

Proposed method

• The system, CourseRank, enables users to rate and review university courses, creating a structured social data repository.
• It employs collaborative filtering techniques to recommend courses based on user preferences and peer reviews.
• The platform supports social features such as user profiles, peer recommendations, and review aggregation.
• It models user behavior and course attributes using a graph-based data model to represent relationships between users, courses, and reviews.
• The system uses a combination of explicit ratings and implicit feedback (e.g., review views, likes) to improve recommendation accuracy.
• It integrates social network analysis to identify influential reviewers and detect potential review spam or bias.

Experimental results

Research questions

• RQ1 Can social systems support more than simple, low-level interactions like 'poking' friends?
• RQ2 What new research challenges do social systems introduce beyond traditional database management systems?
• RQ3 How can social data be structured and managed to support complex, domain-specific applications such as course planning?
• RQ4 What role do collaborative filtering and social network analysis play in improving recommendation quality in academic social systems?
• RQ5 How can social systems be designed to ensure data quality and reduce bias in user-generated content?

Key findings

• CourseRank successfully supports a rich set of collaborative features beyond basic social interactions, including course reviews, ratings, and peer recommendations.
• The system demonstrates that social platforms can be effectively used for structured, domain-specific applications such as academic course planning.
• User-generated reviews and ratings significantly improve the accuracy of course recommendations through collaborative filtering.
• Social network analysis helps identify influential reviewers and detect potential review spam, enhancing data quality.
• The platform reveals that users engage meaningfully with structured social data when the application domain is relevant and useful.
• The study argues that social systems present unique research challenges in data modeling, query processing, and trust management that go beyond traditional DBMS concerns.