A valuable set of inter-related topics are covered in this book. It covers ipython, pandas, cython, numpy, parallel computing, etc. These are all really important parts of the scientific python ecosystem, and there are few or no books out there that properly deal with them all together. For that reason, this book is well recommended. However, I was disappointed in how thin and minimal the coverage of these topic was. Altogether, the book was 119 pages. For this set of topics, this meant that all we got was a brief introduction to each of these topics. That introduction was good, but that's all it was.
Another thing to mention is that this book is titled as if it is a book about IPython. However, it really is not. Ipython is just one of the set of tools it covers. A more appropriate title might be "A good but brief introduction to scientific and numerical python, including ipython, matplotlib, numpy, cython, pandas, parallel python, etc. Essentially, an appetiser".
I guess I am not sure that I really see the point of IPython. I have followed the examples, and get them to work, but I am left asking why I don't just use standard Python in a Visual Studio framework, and simply use the debugging facilities that supplies. I know it is meant to be more of a Matlab replacement, than a Program development environment; but having used both, I am not completely convinced that IPython is all that useful, and hence the need for this book. I will continue to explore IPython, and if I change my mind, I will comeback and revamp this review.
Additional Points. I guess I have to agree with Jean-Luc Ricard's comment about this being more of a statement on Python than a review of the book. I still maintain the 3 star rating. The book is not as well put together as Wes McKinney's book "Python for Data Analysis", but then it is approximately half the length. The book failed to explain to me the advantages of using IPython as against a more conventional approach detailed in the earlier section. Although perhaps this is a purely personal quirk.
Having purchased both this book and the McKinney book; knowing what I know now, I think my recommendation would be if you need to learn IPython, go for the McKinney book.
Many thanks to Jean-Luc for making a very helpful comment.