This chapter introduced us to missing values and underlined the importance of identifying and handling them in order to improve model accuracy.

A number of different mechanisms that underlie the creation of missing attributes were discussed as well as some ways to detect them and then deal with them. The underlying mechanism drives decisions about how to handle the missing values. As with most data exploration techniques, all the situations must be handled on a case-by-case basis, but the importance of missing data means that it is worth having a basic framework to handle it.

Having reached this point, we imported data, cleaned it, and removed outliers and missing values. The next step is to restructure it by transforming it into a different format or by summarizing in new ways to suit the problem to be solved or gain new insights into an overall understanding. This is the subject of the next chapter.