Community urban sensing is one of the emerging applications enabled by the growing popularity of mobile user devices, like smartphones and in-vehicle monitoring systems. Such devices feature sensing and wireless communication capabilities, which enable them to sample large-scale phenomena, like air pollution and vehicular traffic congestion, and upload these data to the Internet. In this work, we focus on the above scenario and investigate the level of accuracy that can be achieved in estimating the phenomenon of interest through a mobile crowdsourcing application. Specifically, we take a signal processing-based approach and leverage results on signal reconstruction from sets of irregularly spaced samples. We apply such results to a realistic scenario where samples are collected by vehicular and pedestrian users, and study the accuracy level of the phenomenon estimation as the penetration rate of the sensing application varies.