Over the last decade, text as data methods have exploded in popularity in the social sciences, digital humanities, and industry. Plentiful data, new methods, and decreasing cost of computation have conspired to make it increasingly possible to unlock the valuable data in documents. These trends toward more data, better methods, and faster computers are only going to increase. Yet, these tools must be adapted to make valid inferences in the social sciences. This book is a guide to harnessing these advancements to make inferences about human behavior both now and in the future.