In the end, sentiment analysis is a simple tool to analyze documents according to two complex, possibly ill-defined categories. Although language is used in complex ways, modern sentiment analysis techniques can do almost as well as humans, which, admittedly, isn't particularly efficient.

What's most powerful about these techniques is that they can provide answers to questions that cannot be answered in other ways. As such, they're an important part of the data analyst's toolbox.

In the next chapter, we'll look at null hypothesis testing, which is a standard and foundational technique of traditional statistics. This informs how we approach many experiments and how we frame the questions that we're asking. By following these guides, we can make sure that our results are more valid and generalizable.