Statistical Techniques - MySQL - page 511

CHAPTER 15
Statistical Techniques
15.0. Introduction
This chapter covers several topics that relate to basic statistical techniques. For the most
part, these recipes build on those described in earlier chapters, such as the summary
techniques discussed in Chapter 8 , and join techniques from Chapter 14 . The examples
here thus show additional ways to apply the material from those chapters. Broadly
speaking, the topics discussed in this chapter include:
• Techniques for characterizing a dataset, such as calculating descriptive statistics,
generating frequency distributions, counting missing values, and calculating least-
squares regressions or correlation coefficients
• Randomization methods, such as how to generate random numbers and apply them
to randomizing a set of rows or to selecting individual items randomly from the
rows
• Techniques for calculating successive-observation differences, cumulative sums,
and running averages.
• Methods for producing rank assignments and generating team standings
Statistics covers such a large and diverse array of topics that this chapter necessarily only
scratches the surface and simply illustrates a few of the potential areas in which MySQL
may be applied to statistical analysis. Note that some statistical measures can be defined
in different ways (for example, do you calculate standard deviation based on n degrees
of freedom, or n -1?). If the definition I use for a given term doesn't match the one you
prefer, adapt the queries or algorithms shown here appropriately.
You can find scripts related to the examples discussed here in the stats directory of the
recipes distribution, and scripts for creating example tables in the tables directory.