Social Networks Privacy Preserving Data Publishing

The proliferation of social networks allowed creating a big quantity of data which contains rich private information that should be preserved. In this paper we consider social networks that are represented as labeled bipartite graphs where each node can have a set of information representing its profile. We propose a solution that allows publishing the social network graphs while preserving the privacy of data. We identify a critical "safety partitioning condition" which has provable guarantees to prevent variety of privacy attacks. We demonstrate the utility of our solution by studying the accuracy with which complex queries can be answered over the anonymized data.