Social Networks Can Also Tell Us a Lot about Non-Members

By May 14, 2012
Keywords : Smart city, Europe
shadow profile

If your friends have registered on a social networking site this means that you could also be identified. This is done by drawing on the email contact lists which are uploaded when a new network user registers.

Some individuals prefer to keep away from social networks for fear that their private information might be revealed. Nevertheless, a study from the Interdisciplinary Scientific Computing centre of the University of Heidelberg in Germany has just offered proof that companies can potentially identify what are known as ‘phantom profiles’, which this study names ‘shadow profiles’. All it takes is for these people to have friends and acquaintances who are themselves registered on the network. The researchers spent several years analysing user profiles on Facebook in order to build their hypothesis. They also looked at previous research which highlighted a correlation between individuals according to the data they shared on the platform – tastes, work, activities in common, etc.

Imported contact list can be revealing

According to previous research, if two or more friends share certain characteristics, there’s a strong chance that a friend they have in common will also share these same characteristics. The researchers wanted to find out whether the theory would also hold good for non-member friends of the users of this kind of platform. So they examined the ‘friend-finder’ application, which imports the user’s email contact list. From this analysis they deduced that if two people with the same characteristics each look out a friend in common by using his/her email address, the friend will most likely share the same characteristics as the other two, whether or not s/he is a member of the platform. The tests and calculations the researchers carried out enabled them to predict 40% of the relationships between non-members.

 Friends of friends of friends…

This very basic knowledge of who is acquainted with whom in the social network can be tied to information about who users know outside the network,” explains Ágnes Horvát, researcher at the University of Heidelberg, adding: "In turn, this association can be used to deduce a substantial portion of relationships between non-members.” This means that individuals outside the social networks can be targeted according to potential profiles identified from the profiles of their friends. The fact remains however that this identification and targeting technique lies at the limit of legality and ethics in relation to data privacy. The researchers were using data drawn from Facebook but without the users’ authorisation so they decided to work with anonymised network data.

Legal mentions © L’Atelier BNP Paribas