Abstract form only given. The large-scale understanding of personal and social behavior from smartphone sensor data is an emerging trend in computing. Smartphones can constantly sense human location, motion, proximity, and communication, and represent one of the most accurate means of tracing human activities. All this data, as never before, is being generated at massive scales. I will present an overview of recent work in my research group in this domain, which includes mobile sensing, data analysis, and applications. I will first describe our experience with the collection of a rich corpus of real-life data using smartphones as sensors, and discuss a few of the many associated challenges. I will then present computational methods that we have developed to discover a variety of patterns, including social interaction types, trends of phone application usage, and personality traits. I will finally discuss about open issues in this domain.