Segaran chooses simple snippets of code to assemble into powerful tools.   The code examples are short and readable, with any Python specific language features fully explained.   By connecting to real data from sites such as del.icio.us, Akismet blogging, and Yahoo Maps, the reader is drawn into the power of these simple techniques to find the hidden power of collective intelligence.
Different facets of collective intelligence are surveyed:
recommendation engines, clustering, searching, reputations, optimized matching, decision trees, and many other classification methods.  The signal to noise ratio is great!  The sheer volume of material covered is staggering!  Each chapter contains explanations and code for several approaches to tricky problems.  In each chapter, a reader is
given useful and solid techniques and is left with an appetite to dig into the references and details.
I would strongly recommend Programming Collective Intelligence, the one with the Penguins on the cover.  It provides serious tools to serious programmers from journeyman to guru.  It is the must-have book
for Web 2.0 and Web 3.0 applications.