Summary

In this chapter, we tuned the performance of the sentiment analysis script from Chapter 9, Analyzing Textual Data and Social Media. Using profiling, Cython, and various improvements, we doubled the execution speed of that example. We also used multiprocessing, Joblib, Jug, and MPI via IPython Parallel to take advantage of parallelization.

This was the last chapter of this book. After the appendices and the index, there is only the back cover. Of course, the learning process will not stop. Change the code to suit your needs. It's always nice to have a private data analysis project, even if it is just for practice. If you can't think of a project, join a competition on http://www.kaggle.com/. They have several competitions with nice prizes. If you are interested in NumPy, you can look forward to the second edition of NumPy Cookbook, Ivan Idris, Packt Publishing, which is planned for 2015.