Inferring Venue Visits from GPS Trajectories

Abstract

Digital location traces can help build insights about how citizens experience their cities, but also offer personalized products and experiences to them. Even as data abound, though, building an accurate picture about citizen whereabouts is not always straightforward, due to noisy or incomplete data. In this paper, we address the following problem: given the GPS trace of a person's trajectory in a city, we aim to infer what venue(s) the person visited along that trajectory, and in doing so, we use honest Foursquare check-ins as groundtruth. To tackle this problem, we address two sub-problems. The first is groundtruthing, where we fuse GPS trajectories with Foursquare check-ins, to derive a collection of detected stops and truthful check-ins. The second sub-problem is designing an inference model that predicts the check-in venue given a stop. We evaluate variants of the model on real data and arrive at a simple and interpretable model with performance comparable to that of Foursquare recommendations.