Uncovering the Location of Twitter Users

Social networks, like Twitter and Facebook, are valuable sources to monitor real-time events, such as earthquakes and epidemics. For this type of surveillance the user´s location is an essential piece of information, but a substantial number of users choose not to disclose their geographical information. However, characteristics of the users´ behavior, such as the friends they associate with and the types of messages published may hint on their spatial location. In this paper, we present a method to infer the spatial location of Twitter users. Unlike the approaches proposed so far, we incorporate two sources of information to learn the geographical position: the text posted by users and their friendship network. We propose a probabilistic approach that jointly models the geographical labels and the Twitter texts of the users organized in the form of a graph representing the friendship network. We use the Markov random field probability model to represent the network and learning is carried out through a Markov chain Monte Carlo simulation technique to approximate the posterior probability distribution of the missing geographical labels. We demonstrate the utility of this model in a large dataset of Twitter users, where the ground truth is the location given by the GPS position, GeoIP location or declared location. The method is evaluated and compared to two baseline algorithms that employ either of these two types of information. The accuracy rates achieved are significantly better than those of the baseline methods.