In order to measure the quality of access to mobile Internet, NIC Chile Research Labs developed Adkintun Mobile, a passive monitor installed in volunteer users smart-phones [2]. Periodically, the client application records data related to network state, which is sent to the collector server. Researchers of the laboratory have access to such stored data. Since from connexion to antennas location can be deduced, location data of individuals is exposed to researchers, which is a concern for location privacy. We propose a model where researchers can access, query and compute aggregations on stored data, learning nothing more about users location than the result of the aggregation. Our model uses homomorphic encryption to preserve location privacy. Location data is sent encrypted from mobile devices to the server. The server can homomorphically evaluate predefined functions such as counting the number of users in a given location (Fig. 1).Query and result decryption are performed from a separate layer, which protects the secret key from being used for direct decryption of the records.