Privacy Risk and Data Utility Assessment on Network Data

In the modern Internet era the usage of social networks such as Twitter, Instagram and Facebook is constantly increasing. The analysis of this type of data can help us understand interesting social phenomena, because these networks intrinsically capture the new nature of user interactions. Unfortunately, social network data may reveal personal and sensitive information about users, leading to privacy violations. In this paper, we propose a study of privacy risk for social network data. In particular, we empirically analyze a set of privacy attacks on social network data by using the privacy risk assessment framework PRUDEnce. After simulating the attacks on real data, we first analyze how the privacy risk is distributed over the whole population. Then, we study the effect of high-risk users sanitization on some common network metrics.