This is a continuation of the books I challenged myself to read to help my career - one a month, for year. You can read my first book review here. The book I chose for Feruary 2012 was: Designing Data Visualizations by Noah Iliinsky and Julie Steele.
Why I chose this book:
As part of the "Data Scientist" track of self-study I've put myself on, I bought the "Data Scientist" bundle from O'Reilly press. As I've mentioned before, this feels less like a coherent, planned offering of a Data Scientist course than a collection of books put together for that purpose. That isn't to say any of these are not good references, because they are, but they certainly weren't designed as a single effort for that purpose. There is no "flow" of information from one to the other, for instance.
Some of these works aren't as useful as others, at least to me. For one thing, they tend to vary remarkably in depth. One book dives deeply into statistics, assuming that you have a background knowledge in undefined knowledge-sets, and another starts at the beginning. This book falls into the latter category - but more on that in a moment.
Even so, I chose the collection because of the price, and the range of information that they contain. I feel that the Data Scientist role deals with four areas:
1. Data Sourcing and Acquisition
2. Storage and Processing
This particular book falls into the last category, and since I desire more artistic talent than I have, I felt it would be quite useful. I'm always looking for information that helps me formulaicly look like I have an artist's eye. I'm quite envious of people who can draw and make awesome graphics. And at the end of the day, information needs to be presented in the best way to expose it to the right audience. Yes, before you ask, I have read Tufte's work. Some of his work is good; other parts of it, not so much.
What I learned:
I actually read this entire book on a single plane ride. It's a small book, and allows you to grok the information quickly. It's put together for the beginner, but I don't think it "talks down to you" or makes you feel that it's too basic.
When I began the book, I had a different expectation - I thought there would be a lot of theory, that sort of thing. It's actually quite good as a reference - it goes through some of the best ways to present information, has a quick chart on when to use which kind of visualization, and has a fantastic reference section at the end. In fact, the reference section alone is a large part of the value of the book.
The author does stray a bit into general design concepts, and talks more about web presence and so on than I think is needed for a pure Data Scientist. Industry opinion is still out on whether a Data Scientist needs to do the final visualizations or not, some feel that this is the job of the Business Intelligence professional. Personally, I think presentation is part of life in general, and so this book is at least useful in that way for someone who needs a basic introduction.
And I suppose that's how I'd sum up this book - a great introduction to data visualization that you can read quickly. I recommend it for your library if you work with data in your profession.