Methods and systems for characterizing mobile parking sensor data. With a generative model, a targeted random variable can be identified from parking data collected from one or more mobile sensors. The parking data includes data indicative of time instants and geographical locations. In some cases, a proportion of parking stalls associated with payment data can be derived from the targeted random variable at each time instant and for each street block face among the geographical locations. Parameters of the generative model can be determined base on observed data that is at least partial in time. The generative model is then applied in order to infer quantities of interest for use in characterizing the parking data including a price and a rate with respect to the quantities of interest.
