The capability to anticipate a contact with another device can contribute to improving the performance and user satisfaction of mobile social network applications and of any other relying on some form of data harvesting or hoarding. This paper presents a nine year data set of wireless access logs produced by more than 70,000 devices and 40,000 users. Research on the recurring contact patterns observed between groups of devices permitted to model the probabilities of occurrence of a contact at a predefined date between pairs of devices. As an example, the paper presents and evaluates an algorithm that provides daily contact predictions, based on the history of past pairwise contacts and its application on a reputation service.