3G Smartphone Technologies for Generating Personal Social Network Contact Distributions and Graphs

This paper presents a novel means of collecting and analyzing data related to personal social contact networks. The work developed a custom application for Smart phones that support Bluetooth connectivity, as representative of the ensemble of many consumer electronic products and can be used to infer users´ proximity, contact duration, and GPS-based information. In many cases this is augmented by device meta identity. The 3G application, data storage and retrieval is discussed in detail. Preliminary data were collected (device-device proximity, duration, and location) gathered in pilot testing on the Blackberry Storm and HTC Hero (Android). Data are presented as distributions and visualization tools for evolving contact graphs, including Pareto distributions and power law exponents generated representative of face to face contacts. Finally, data are then demonstrated to be useful in estimating the potential of infection spread (e.g. respiratory illness), where a key transmission vector is person-person contact. A variant of the standard SIR agent based model is developed, with agent contacts guided by contact distributions extracted from the data.