This paper shows that human personality can be accurately predicted by looking at the data generated by our smartphones. GPS location, calls, battery usage and charging, networking context like bluetooth devices and WiFi access points in proximity, and more give enough information about individual habits, reactions, and idiosyncrasies to make it possible to infer the psychological traits of the user. We demonstrate this by using machine learning techniques on a dataset of 55 volunteers who took a psychological test and allowed continuous collection of data from their smartphones for a time span of up to three years. Openness, Conscientiousness, Extroversion, Agreeableness, and Neuroticism (the so called Big5 personality traits) can be predicted with good accuracy even by using just a handful of features. The possible applications of our findings go from network optimization, to personal advertising, and to the detection of mental instability and social hardship in cities and neighborhoods. We also discuss the ethical concerns of our work, its privacy implications, and ways to tradeoff privacy and benefits.