GeoSRS : A Hybrid Social Recommender System for Geolocated Data Joan Capdevila a,c,1 , Marta Arias b,c,2 , Argimiro Arratia b,c,2,3 a Barcelona Supercomputing Center b Computer Science Department c Barcelona Tech / Universitat Polit` ecnica de Catalunya, Spain Abstract We present GeoSRS, a hybrid recommender system for a popular location- based social network (LBSN), in which users are able to write short reviews on the places of interest they visit. Using state-of-the-art text mining tech- niques, our system recommends locations to users using as source the whole set of text reviews in addition to their geographical location. To evaluate our system, we have collected our own datasets by crawling the social network Foursquare. To do this efficiently, we propose the use of a parallel version of the Quadtree technique, which may be applicable to crawling/exploring other spatially distributed sources. Finally, we study the performance of GeoSRS on our collected dataset and conclude that by combining sentiment analysis and text modelling, GeoSRS generates more accurate recommenda- tions. The performance of the system improves as more reviews are available, which further motivates the use of large-scale crawling techniques such as the Quadtree. Key words: recommender systems, text mining, quadtree, crawling, social networks, location-based social network Email addresses: jc@ac.upc.edu (Joan Capdevila), marias@cs.upc.edu (Marta Arias), argimiro@cs.upc.edu (Argimiro Arratia) 1 Supported by Obra Social “la Caixa” 2 Supported by MICINN project TIN2011-27479-C04-03 (BASMATI), MINECO project TIN2014-57226-P (APCOM) and Gen. Cat. project SGR2014-890 (MACDA) 3 Additional support by MEC project MTM2012-36917-C03-03 (SINGACOM) Preprint submitted to Mining Urban Data October 30, 2015