This is the most technical chapter of the book. Metric selection is discussed in some detail before moving on to new visualization techniques to represent complex spatial processes evolving over time. It then digs deeper into the technical aspects of a range of topics.
This chapter presents some interesting techniques developed recently to handle modern business data. It is more technical than previous chapters, yet it's easy to read for someone with limited statistical, mathematical, or computer science knowledge. The selection of topics for this chapter was based on the number of analytics practitioners that found related articles useful when they were first published at Data Science Central. This chapter is as broad as possible, covering many different techniques and concepts.
Advanced statistics, in particular cross-validation, robust statistics, and experimental design, are all part of data science. Likewise, computational complexity is part of data science when it applies to algorithms used to process modern large, unstructured, streaming data sets. Therefore, many computer science and statistical techniques are discussed at a high level in this chapter.
NOTE Material that is traditionally found in statistical textbooks (such as the general linear model), as well as run-of-the mill, old computer science concepts (such as sorting algorithms) are not presented in this book. You can check out Chapter 8, “Data Science Resources,” for ...