SniffMob: inferring human contact patterns using wireless devices.

ABSTRACTThe size of existing data sets regarding human mobility and person-to-person contact has been limited by the labor-intensive nature of the data collection techniques employed. In this paper, we propose a practical data collection system which is automatic and transparent to the user, requires only installing new software, and uses the multiple sensing capabilities provided by current commodity mobile devices. This approach allows the scale and duration of these human contact studies to increase by several orders of magnitude and allows for the collection of location and contact information about individuals who do not install our data collection software. We present an analysis of the expected coverage of our data collection system drawing from existing data sets and random graph theory. To illustrate the type of application enabled by the availability of human contact data, we present "personalized epidemiology," a novel application that provides its users with information about their exposure to illness and offers advice on how to remain healthy.