Using mobile phone global positioning systems (GPSs), location data for individuals can be gathered and analyzed for various purposes. Using two-level clustering, Montoliu and Gatica-Perez gather this data for the purpose of investigating individuals’ places of interest. Data for eight people over a period of five months revealed their places of interest. About 63 percent of their daily locations were captured.

In this setting, the subjects used phones with GPSs--there was no need for a beacon location database. The authors introduce an algorithm that clusters the GPS data into stay points and stay regions. The geographical and temporal data was analyzed in a client-server mobile application. Places can be discovered, remembered, and forgotten--the algorithm takes care of that. Also, parameters can be defined in the grid-based clustering technique to make the algorithm perform on other data points. The places of interest discovered were juxtaposed against user statements. The so-called experiments are testimonies of tasks completed within the data collection and data crunching process. Compared to two similar methods, this approach seems to perform the best.

The paper is easy to read. The idea and method are easy to understand and follow. Perhaps the next step is implementation in a sensory system context or a geo-interest discovery social application.