Are you moved by your social network application?

This paper studies a Bluetooth-based mobile social network application deployed among a group of 28 participants collected during a computer communication conference. We compare the social graph containing friends, as defined by participants, to the contact graph, that is the temporal network created by opportunistic contacts as owners of devices move and come into communication range. Our contribution is twofold: first, we prove that most properties of nodes, links, and paths correlate among the social and contact graphs. Second, we describe how the structure of the social graph helps build forwarding paths in the contact graph, allowing two nodes to communicate over time using opportunistic contacts and intermediate nodes. Efficient paths can be built using only pairs of nodes that are socially close (i.e. connected through a few pairs of friends). Our results indicate that opportunistic forwarding complies with the requirement of social network application.