Vincent Granville

There are some good nuggets of information here and in many cases way too much information. But then maybe that is the hazard of being a data scientist/author--you want to give everybody ALL the information so that they can then dig out the nuggets themselves. Unfortunately that does not make a good technical book. IMPO (in my professional opinion as a former editor for technical books) a good technical book is an author taking the reader by the hand and telling them what they need to know--not spewing gads of links, bulleted lists and drifting off topic in the middle of the chapter to explore a math problem. I could blame the editor or maybe this was just a cowboy author. Obviously the author is brilliant--but that doesn't make this a good or easy book to read.… (más)

[Leaving aside the, by now, well-known, pay for Amazon reviews episode.]
This is a mess of a book. Based on other reviews, I see I'm not the only one who found it confusing and certainly not a good illustration of the title. A big chunk of the book is dedicated to defining data science as opposed to what the author calls "fake data science". That is done in a tone that goes from condescending to resentful (with the usual critique that [insert your favorite emerging field] is not properly taught in university, that's par for the course). Then, the author abruptly goes into illustrations that people with no background in statistics (but not the old, bad, statistics, the new cool one developed by the author) will have a hard time following (so, so much for the "developing" part in the title). And the book ends with some job market data (interview questions, salary survey, etc.).
In the end, the author does not give a lot of info on how to start in data science. The entire book, and especially the case chapters, assume extensive knowledge of the field. Readers will find bits and pieces of useful information here and there, but it feels like needles in a hay stack.

… (más)