Our research focuses on methods for extracting latent contexts from raw data, and utilization of latent context for
context aware recommender systems. As a case study, we analyzed data obtained from mobile device sensors and compared the recommendation quality of our approaches with other state-of-the-art context aware recommender systems. We demonstrate the contribution on a system that recommends points of interest (POIs) retrieved from Foursquare 1 API, e.g., restaurants, bars, cafés, entertainment centers, etc. The idea is to model the context for which the POIs are relevant and to use the context in the recommendation process, along with Algorithm 3 described above. In order to obtain data for the evaluation of the algorithm, we developed an Android application which monitors the user’s location and recommends popular POIs nearby.