This chapter has introduced some more advanced techniques for transforming data into a format more suited to exploration and data mining. These included generating attributes based on other attributes in the same example as well as attributes in other examples through the use of macros. In addition to this, aggregation, pivoting, de-pivoting, and windowing were all discussed.

In my experience, a great deal of time is spent transforming data using these techniques, and it is worthwhile investing time learning how to use them.

The next chapter considers how to reduce data size by simple sampling methods or more complex model-based approaches. While on the face of it this may seem like a bad idea, the reality of real data is that there can be too much of it and it would take too long to process unless some reductions are done.