ASA members Bin Yu of UC Berkeley and Rebecca Barter of Utah University Medical School recently released the open-access version of their new book, Veridical Data Science: The Practice of Responsible Data Analysis and Decision Making.

The book demonstrates how to use the principles of predictability, computability, and stability to create and evaluate trustworthy data-driven results by conducting stress tests at every stage of the data science life cycle—from problem formulation and data cleaning to modeling and communicating results. 

Veridical Data Science arose from the graduate statistics class Yu has taught for many years at Berkeley and has evolved to embody her research philosophy. It is a resource for data science, statistics, and applied machine learning classes as one of the textbooks or supplementary reading. It is intended for upper-division and beginning graduate students, as well as for domain scientists entering data science.

The print version will be available late this year through MIT Press.