Crowdsensing applications are increasing at a tremendous rate. In crowdsensing, mobile sensors (humans, vehicle-mounted sensors, etc.) generate streams of information that is used for inferring high-level phenomena of interest (e.g, traffic jams, air pollution). Unlike traditional sensor network data, crowdsensed data has a highly skewed spatio-temporal distribution caused largely due to the mobility of sensors . Thus, designing systems that can mitigate this effect by acquiring crowdsensed at a fixed spatio-temporal rate are needed. In this paper we propose using multi-dimensional point processes (MDPPs), a mathematical modeling tool that can be effectively used for performing this data acquisition task.