Proceedings of the 2nd ACM SIGMOD Workshop on Databases and Social Networks

The ACM SIGMOD Workshop on Databases and Social Networks (DBSocial) aims at disseminating results founded on database research and practice that advance the state-of-the-art in the observation, management, and analysis of inherently networked data that results primarily from social phenomena. In particular, DBSocial is intended to foster a discussion about the role that the database community should play in the area of social network research. As such, DBSocial welcomes papers whose approaches are fundamentally centered on theoretical foundations and best practices in databases and very closely-related areas such as data mining and information retrieval. In its second edition, DBSocial features three invited talks and five peer-reviewed papers (selected out of twelve submissions). The invited talks cover challenges and advances in building the infrastructure necessary for storing, accessing, and analyzing large social networks, as well as supporting applications that use them. Amol Desphande from the University of Maryland talks about challenges in building a scalable data management and analytics tool capable of handling large networks, and some advances in his group towards that end. Pankaj Gupta and Jimmy Lin discuss components of the Twitter infrastructure for handling large and sparse networks of users. Finally, Adam Silberstein gives the perspective of LinkedIn, detailing components in their infrastructure, and how they are used to back their diverse set of social applications. The papers cover a wide range of topics related to DBSocial. A technique for solving race conditions in a multi-tiered database system similar to those used in social networking sites is proposed by Ghandeharizadeh and Yap. Next, Ahmed and Guha present an exploratory analysis of how linguistic features correlate with geographical information about users, based on a dataset from Flickr. The paper by Jouili, O-Doherty and Van Roy discusses the problem of trust inference between users based on common preferences, which remains a challenging problem as users seldom explicitly indicate trust. Xiao, Tan, and Aung consider the problem of managing "circles" of connections in social networks, which is an emerging problem arising from the popularity of social networking sites. Closing the proceedings, the paper by Ribeiro and Silva proposes an algorithm for finding the occurrences of a set of graphs within in one graph. Their approach is to use indexing structures called G-Tries, which are shown to lead to substantial efficiency gains.