Book Review: Data Science for Business

Next up in my 3-post book review series is Data Science for Business. When I sifted through Amazon for highly rated data science books over the holidays, I wanted one that was targeted at a general audience to gain some perspective on what non-techies are reading about the field. I stumbled across several novel-sized softcovers like Nate Silver's "The Signal and the Noise" and "Naked Statistics: Stripping the Dread from the Data", but I find most books on technical subjects made to look like novels read like novels: they prioritize anecdotes, interesting factoids and wittiness over substance. Data science for business doesn't get too deep into math or programming, but it manages to provide a business-minded overview of data science with enough substance to give readers a solid understanding of what many common data science methods do and how they are used.

Data Science for Business covers a range of machine learning concepts and techniques including the data mining process, linear models, tree-based models, overfitting and assessing model performance. The words "for business" shape all of the content. Although data science is the the subject, the topics are always discussed in the context of business. For instance, the section on assessing model performance focuses on expected profit rather than standard measures like accuracy, sensitivity, specificity, Kappa or AUC. The book uses case studies and examples where appropriate to show how the concepts work in practice, but it doesn't go overboard with anecdotes and other flavorful fluff. Given the title, the book actually gets a little more technical than I had expected, although it is written in such a way that readers can skip the most difficult sections. The book starts off a bit slow and if you've studied data science, a lot of it will be review. On the plus side, the reading goes rather quickly because it doesn't bogged down in equations and low level details.

Data Science for Business is worth a read even if you already know the basics of data science, as it gives insight into what people outside the field are reading about it and it has enough substance that you'll probably learn some new things in the process.

I give Data Science for Business a rating of 4.5 out of 5 stars: Great.