The Art of Statistics: Learning from Data

Abstract

At the outset of the book, the author notes two modern challenges to the traditional view of statistics as a basic ‘bag of tools’: the inappropriate application of statistical methods, resulting in the crisis of reproducibility among scientific publications, and the emergence of data science, in which large datasets are analysed by increasingly sophisticated algorithms but with diminished human oversight. As the Winton Professor of the Public Understanding of Risk at the University of Cambridge and former President of the Royal Statistical Society, David Spiegelhalter is extremely well placed to offer a remedy to these developments. To this end, Spiegelhalter has produced a highly readable and engaging introduction to an impressive array of statistical concepts, emphasizing common pitfalls at each step. Eager to avoid the litany of mathematical formulae which often pepper introductory statistical texts, Spiegelhalter motivates each concept with compelling questions derived from real-world case studies, such as calculating...