Data Mining Book Review: Naked Statistics
I heard about Naked Statistics, from Charles Wheelan, a few months ago. Marcel Baumgartner mentioned it on the LinkedIn group of the Swiss Statistical Society as an excellent book. So I urged to bought it…and I wasn’t disappointed. Naked Statistics is a journey in the world of statistics applied to everyday challenges.
The book is very readable and is often funny. Wheelan introduces some concepts through a new TV show: CSI Regression Analysis. Each statistical concept is presented through several examples. Topics covered range from descriptive statistics to correlation/causation and false positive vs. false negative, among others. The Monty Hall problem is also very well explained.
The book goes on with more advanced topics such as randomized experiments. The author explains several challenges related to such methods and how to surpass them. Wheelan also discusses hypothesis tests and issues related to confidence level and p-value. The book is an homage to How to Lie with Statistics and it is complementary to it. Naked Statistics is a delicious surprise that will delight anybody interested in statistics and analytics.