Matplotlib is a popular 2-dimensional plotting library used with Python. While Matplotlib is fairly powerful, I have always had trouble figuring out how to do what I needed from the documentation. I have found the online Matplotlib documentation adequate as a reference, but not very good for getting started (especially since Matplotlib uses some terms in what I would call a "non-standard" way).
Matplotlib for Python Developers by Sandro Tosi is a welcome addition to my bookshelf. It focuses on embedding Matplotlib in applications in GTK+, Qt 4, wxWidgets, and even various web frameworks such as Pylons and Django. But the fundamental elements of the scientific Python tool chain, such as NumPy and IPython, as well as a thorough tutorial of Matplotlib itself are discussed. Both the pylab/pyplot procedural interface (useful for interactive plotting) and object-oriented interface (useful for fine-tuning and more advanced usage) are covered. Over 2000 lines of example code are downloadable from the book site.
Seeing as I don't use Matplotlib regularly, I'm sure I will using this book frequently to brush up on Matplotlib basics. And this book would have saved me a lot of time when I was writing my only major Python project, a PyQt 4 application which embedded interactive Matplotlib graphics. While this book doesn't cover every detail of Matplotlib (the online docs are good for that), it will get you started no matter how you are using it.
The publisher has made a sample chapter on embedding Matplotlib in Qt4 available for download. More information, such as a full table of contents, is also available on the publisher's site.
Full disclosure: I received a review copy of this book.