- ISBN: 978-1-449-30711
- Publisher: O’Reilly
- Publication Date: July 2011
- Pages: 119
- Price: £22.99
Think Stats is designed to teach probability and statistics to programmers through hands-on examples. The language of choice is Python, and throughout the textbook exercises are assigned to analyse real world data through the use of Python programs.
The example code provided is not always explained very thoroughly, but Python is fairly intuitive, so this should not put off programmers specialising in other languages. The book is, however, very short (120 pages) and I feel that it could have been made longer by adding some detail into example code and some of the explanations. Few assumptions of prior knowledge of statistics is made, however many of the details, some of which can be quite complex, are glossed over. Downey instead refers readers to other references; many of which are Wikipedia. He references Wikipedia many times as it is a source of free information which should be accessible to all, unlike certain journals. This need of constant external sources does however break up the reading experience, negatively impacting my impression of the book. Likewise, example code is provided as solutions to some of the exercises, but it must be downloaded from the book’s website, rather than having key snippets included in text.
The overall tone of the book was very informal, which personally I liked. It’s hard to introduce a sense of joviality into a book dealing with statistics, but I found myself laughing out loud on several occasions.
The glossary provided at the end of each chapter is invaluable and is a great addition to the book, but it does very little to mitigate the high price. Think Stats is actually released under a Creative Commons Attribution-NonCommercial 3.0 Unported license. This means that it is free to be copied and distributed for free. In fact, you can read it online or download a PDF from the book’s website; thinkstats.com. Considering that it is available for free, the short length, and that not every exercise has a solution provided, I find the price of £23 to be far too high. For that price I would have liked to have seen at least a companion CD with code listings and a Python compiler included.
Think Stats is a good primer on statistics and probabilities, but its high price makes it hard to recommend for your bookshelf. Instead, go and read it for free and you won’t be disappointed.