STEPS: Predicting place attributes via spatio-temporal analysis

In recent years, a vast amount of research has been conducted on learning people’s interests from their actions. Yet their collective actions also allow us to learn something about the world, in particular, infer attributes of places people visit or interact with. Imagine classifying whether a hotel has a gym or a swimming pool without ever talking to the people who stayed there. Or imagine predicting whether a restaurant has a happy hour or a romantic atmosphere without asking its patrons. Algorithms we present can do just that. Many web applications rely on knowing attributes of places, for instance, whether a particular restaurant has WiFi or offers outdoor seating. Such data can be used to support a range of user experiences, from explicit query-driven search to proactively recommending places the user might like. However, obtaining these attributes is generally difficult, with existing approaches relying on crowdsourcing or parsing online reviews, both of which are noisy, biased, and have limited coverage. Here we present a novel approach to classifying place attributes, which learns from patrons’ visit patterns based on anonymous observational data. Our method, STEPS, learns from aggregated sequences of place visits. For example, if many people visit the restaurant on a Saturday evening, coming from a luxury hotel or theater, and stay for a long time, then this restaurant is more likely to have a romantic atmosphere. On the other hand, if most people visit the restaurant on weekdays, coming from work or a grocery store, then the restaurant is less likely to be romantic. We show that such transition features are highly predictive of place attributes. We also introduce a variant of our method, STEPS-E, which builds a highdimensional embedding model trained on co-visitation data, and completely eliminates the need for feature engineering. In an extensive empirical evaluation, STEPS nearly doubled the coverage of a state of the art approach thanks to learning ∗Corresponding author ISBN X-XXXXX-XX-X/XX/XX. DOI: from observational location data, which allowed our method to reason about many more places.