Book Review: 21 Recipes for Mining Twitter

“21 Recipes for Mining Twitter” by Matthew A. Russell is written for a technical audience and specifically programmers that use or are comfortable with the Python programming language.  The bulk of the book is a collection of Python code samples (a.k.a., recipes), and suggested tools, that will help you develop the code that you need to extract data from Twitter and present this information for analysis.

Matthew Russell presents a clear description of the problem that each recipe solves and discusses some of the finer points of the recipe before diving into the code.  This concise writing style makes the concepts easy to understand, which begin with more basic concepts such as using OAuth to access Twitter APIs to more complex topics such as visualizing geodata with a Dorling cartogram.

The author also provides useful recommendations for tools to utilize in conjunction with the data that you are able to collect from Twitter.  For example, using Protovis (JavaScript visualization toolkit) or the WP-Cumulus WordPress plugin for web visualization, or Apache CouchDB to natively store JSON data.

One substantial benefit of reading this relatively short book–it is only 72 pages in length–is that it identifies and will therefore help you avoid common pitfalls.  For example, one section focuses on writing robust requests that can handle the more common issues, such as Twitter’s rate enforcement.