Data Mining & Statistical Analysis Using SQL

Data Mining and Statistical Analysis Using SQL concerns itself with the interface between applied mathematics--the discipline of statistical analysis--and really applied mathematics in the form of Structured Query Language (SQL) code that carries out such analysis. It's a subject that deserves careful coverage in a book, and the authors of this one--both working analysts with distinguished academic backgrounds--have done great work. If you're faced with a need to derive meaning from large quantities of data (from retail sales, industrial processes, or even scientific observations), and canned analysis tools aren't cutting it for you, take time to study what Robert Trueblood and John Lovett have to say.

Though some background in statistics will help you pick up on what Trueblood and Lovett have written, a low-level university class (even one far in your past) should be enough. Their approach to all of the analysis techniques they teach is to explain terms and concepts with prose, then with graphs, then with formulas. Then, they translate the formulas into SQL queries for Microsoft Access and show variations on the code that yield differently tweaked results. Finally, T-SQL source code (for Microsoft SQL Server 2000) is listed, though most readers will prefer to grab this code from the book's companion Web site. Additional coverage of graphics would make this book better, but in its present state it's great reading for people who want to interpret their mountains of data. --David Wall

Topics covered: Statistical analysis as a set of mathematical tools that may be implemented in Structured Query Language (SQL), specifically SQL variants for Microsoft database products. Chapters explain how to use hypothesis testing, curve fitting, scatter plots, measurements of central tendency, and regression analysis to spot significant characteristics of data.