For years, researchers and network providers have been interested in the spatio-temporal distribution of data rates in mobile networks. However, data rates are often studied by averaging over long time intervals and large cell sizes, concealing important properties, which are needed for physical layer design approaches that aim to exploit the variability of data rates within a cellular network, e.g., for energy savings. Moreover, recent research has highlighted a high spatial and temporal correlation between data rates of mobile social and cellular networks, which motivates the development of a common framework. To address these issues, this paper proposes a high-time-resolution model for data rates in a mobile network, with a parameterization that presumably makes it adaptable to both cellular and mobile social networks. The model incorporates spatial and temporal correlation and enables the derivation of the probability distribution of data rates occurring in mobile networks. We fit the model with a data set of Twitter tweets and compare the data rates predicted by the model with measured data.