Analytics of Social Sensing

Abstract: The vision of smarter cities that better conserve their resources and better streamline their services relies in part on the increased availability of data about the real time state of such resources and services, and the increased ability to perform large-scale data analytics. A central architectural component is therefore an infrastructure for data collection and information distillation that relies not only on fixed sensors but more importantly on people as data sources. Social sites, such as Twitter, allow individuals to share large amounts of data for social sensing applications. This talk discusses the research challenges brought forth by the need to collect reliable data in real time from large groups of individuals who may be unknown, unreliable, or not motivated to collect and share their data. Mathematical formulation of the underlying problems are presented together with analytic solutions. Initial evaluation results are shown from experimental prototypes deployed in controlled settings.