In this O'Reilly report, Jerry Overton--Distinguished Engineer at global IT leader CSC--introduces practices for making good decisions in a messy and complicated world. What he simply calls ''data science that works'' is a trial-and-error process of creating and testing hypotheses, gathering evidence, and drawing conclusions. These skills are far more useful for practicing data scientists than, say, mastering the details of a machine-learning algorithm. Adapted and expanded from a series of articles Overton published on O'Reilly Radar and on the CSC Blog, each chapter is ideal for current and aspiring data scientists who want to go pro, as well as IT execs and managers looking to hire in this field.