This is a review I completly agree with:
Mining the Social Web is a good start for anyone is going to create scripts to analyze patterns in Social Networks. I've to say that this book consider that the reader already masters Python. I think that should be written directly on the title (ie: "Mining the Social Web with Python").
I liked the really fast approach to the Social Networks, even if a lot of times I wanted more; for this reason I consider it only a start, not a complete book.
The part I really enjoyed was the one about the HTML5 microformats. This is the only book that cover the topic from the data mining point of view, as far as I've seen.
I had the pleasure to read this book in the e-book version, and I've to admit that O'Reilly did a really good job in linking the different parts of the book through hyperlinks.
The biggest part of the book is focused more on what text is important rather that why it is important. In the book is touched the why speaking about the Semantic Web processing, but this is beyond the goal of this book, I suppose.
There are also a number of areas out of date. Twitter API data has been changed considerably and these are covered in the errata on the website. In other cases, a whole chapter is dedicated to Google Buzz which no longer exists. But it doesn't matter since the most interesting things are the techniques explained in the whole book. Another aspect I appreciated was that the code examples were all in a github repo. I found some performance / memory issue in one of those so I forked the repo , did my patch and submitted a pull request that has been accepted! The author of the book honored me with a very kind comment:
It's a nice and elegant solution. Wish I'd have thought of it myself :)So fun, profit and satisfaction!