Gartner, an information technology (IT) research organization, tracks emerging and maturing technologies using its famous “hype cycle” graphics [1]. Data science, big data, and related analytics have followed Gartner’s iconic sequence of innovation, inflated expectations, disillusionment, enlightenment, and productivity, seemingly very quickly in the past few years. Data science is now a mainstream career area, as well as a major focus in industry, government, and academia, with numerous domain-specific applications in medicine, social media, and commerce. Scientific research and engineering are no exception to this evolutionary trend, and are now clearly in the hype cycle’s productivity phase.

Brunton and Kutz present a comprehensive and detailed textbook on applying modern mathematical and programming data analysis methods to science and engineering research. They are well-known researchers at the University of Washington, as well as data science fellows at the school’s eScience Institute. Both have won prestigious teaching awards.

The book is intended for advanced undergraduate and graduate engineering and physical science students with strong backgrounds in linear algebra, differential equations, and scientific computing, who need instruction in modern data analytics methods for optimization and control systems, with a focus on large datasets and machine learning. It opens with a review of common optimization techniques, symbols, and equations, and then covers four major areas: transformations, machine learning, control systems, and reduced order models.

The presentations are heavily mathematical yet succinctly and understandably explained, covering basic concepts using classic use cases and applications emphasizing dynamic systems. The section on machine learning, covering regression and model selection, clustering and classification, and neural networks and deep learning, is particularly illuminating and instructive.

As might be expected from a Cambridge University Press publication, the textbook is quite attractive, with high-quality printing, illustrations, and color graphics, along with helpful suggested readings and references at the end of each chapter, a thorough and informative glossary, an extensive bibliography, and a companion website (http://databookuw.com/) that includes outstanding YouTube video lectures, homework problems using MATLAB and Python, and numerous illustrative datasets. The index is a bit sparse, missing some key items from the text, but the book can still be useful as a reference for the numerous analytics concepts and methods, especially when paired with the accompanying website.

Overall, this valuable and accessible textbook for its intended audience spans introductory material through complex modern analytics practices. Students who master its content will be well prepared when it comes to encountering today’s challenging problems.

More reviews about this item: Amazon