The fluff, reviewers comment, and marketing hype around this book include such revelations as “Load important statistical concepts directly into your brain” and “Wouldn’t it be dreamy if there was a book on data analysis that wasn’t just a glorified printout of Microsoft Excel help files?” However, this volume is just a little bit different from what you may be used to in an intro book or “Dummies Guide.” Employing real life examples, an Excel spreadsheet, and the R programming language, the author attempts to structure problem solving in a way that will be exciting and stick to (or in) your brain.

A quick look at the Table of Contents will tip the reader off that this book takes a different approach to teaching data analysis (or is it leading by example?). We are offered the standard fare such as hypothesis testing, experimental design, regression and Bayes’ Rule but in such a way as to avoid the usual pedantic and unimaginative (read boring) approach that many books take. Rather than introducing t-tests and analysis of variance as dry and fixed obstacles to be overcome, the author uses Head First learning principles such as “make it visual,” “use a conversational style,” “keep the readers’ attention” and “touch their emotions.”Although heavily integrating graphics into the analysis is hardly a new technique, the author’s use of simple, sparse graphs to clarify what the data is saying is quite refreshing.

From my point of view, the data analysis is of course valuable, but the tips on learning are priceless. The following points are listed in the introduction under “Here’s what YOU can do to bend your brain into submission:”

1. Slow down. The more you understand, the less you have to memorize.
2. Do the exercises. Write your own notes.
3. Read “There are No Dumb Questions”
4. Make this the last thing you read before bed, or at least the last challenging thing.
5. Talk about it — out loud.
6. Drink water — lots of it.
7. Listen to your brain
8. Feel something.
9. Get your hands dirty.

Each of these are explained or justified in a few brief sentences. For example, the last item refers to the fact that far more learning will take place if you actually do the exercises. I find many of these to be valuable tips and not just motivational lecture! The author stresses that “this is a learning experience, not a reference book.” The other core detail that can be stressed is that this is about data analysis and not just statistics. This is not a trivial distinction as in data analysis we are focused on much more than just the analytic concerns, e.g. correctly defining the problem, reducing the problem to manageable bites, knowing the client and making decisions based upon the analysis. Some statisticians may be doing all these steps in their daily activities, but they are not always equivalent.

The real value of this book lies not in its meager statistical content (and I say that as a statistician) but in its approach and ancillary content. Readers probably already know that they need to explicitly and concisely define a problem and that they need good information to solve it, but the book introduces the many curveballs and sidetracks that occur in real, everyday problem solving that may be easily glanced over. By emphasizing these pitfalls and stressing the ways to address them, as well as maximizing the learning experience, it delivers far more than a simple problem solving technique. A quick reading (Oops! The author tells us not to go too fast when learning) of the first chapter illustrates this, as the reader quickly sees that the original data and client perceptions can mislead the analyst. Sometimes it is necessary to dig deeper before we attempt to think deeper.

For those readers new to data analysis and/or easily intimidated by introductory statistics, this is highly recommended. Give a copy to your manager (but only if he or she falls into the preceding categories)!

Availability
Head First Data Analysis, by Michael Milton. O’Reilly Media, Inc. Sebastopol, CA. 445 + xxxviii pp. 2009. $49.99

John Wass is a statistician based in Chicago, IL. He may be reached at editor@ScientificComputing.com.