mining the social web - matthew russell

Mining the Social Web does a great job of introducing a wide variety of techniques and wealth of resources for exploring freely available social data and personal information. If you are willing to spend the time tinkering with the examples, the book is pure fun. It offers a nice compliment to Segaran's Programming Collective Intelligence. The two books overlap but where they do offer different perspectives and explanations of common techniques (e.g., TF-IDF, cosine similarity, Jaccard index). If you are well-versed in data mining the web you may find much of the discussion familiar. If you have only been casually engaged to date, your toolbox will fill quickly.

In order to work with the book's examples related to LinkedIn and Facebook you really need to have a robust collection of connections. In terms of the source code itself, most of it worked as is. I wasn't able to install the Buzz library which limited my interaction with material in chapter 7 and opted to not get involved with the LinkedIn or Facebook but found the discussions around them easy to follow. By far my favorite chapter in the book was chapter 8, "Blogs et al.: Natural Language Processing (and Beyond)..." It was quite fascinating and caused my reading list to grow considerably.

Full information: Mining the Social Web, Matthew Russell, O'Reilly Media, Inc.