An integrated platform for collecting mobile phone data and learning demographic features

The problem of collecting, processing, and learning from high-volume mobile device data has become an active research area in recent years. Time series data on application usage, in particular promises to provide fine-grained information on individual activity patterns, but currently poses collection and analysis challenges. In this paper we demonstrate an integrated system which can cheaply and easily collect application behavior and survey data from mobile phones; we introduce several novel features that assist the learning of individual level demographic features (e.g., gender and age group). Specifically, our approach for learning and inference for demographic features involves new techniques: (i) decomposing the app usage from mobile phones using spectral methods; (ii) learning spectral characteristics associated with individuals using a training set; (iii) combining other temporal features with learned spectral characteristics to predict demographic features for out-of-sample individuals. The core of our methodology is the utilization of spectral features in cell phone app activity series, allowing both identification of behavior patterns arising from particular types of cell phone apps and leveraging of those patterns for demographic classification and prediction. We demonstrate the effectiveness of our approach with an application to real mobile app traffic data from the United States.