Desktop Survival Guide
by Graham Williams
Reporting is an important part of any data mining project. A report will generally capture a summary of the business problem, the sources of the data, the data processing performed, the data mining performed, and the results. Following the philosophy of literate programming (perhaps we might call it literate data mining), an ideal report will include results generated directly from the data mining process, as well as each of the steps involved and the actual code (which may or may not be included in the resulting processed document). The Sweave and odfWeave packages of R support literate data mining for LATEX and OpenOffice.org, respectively.
In this chapter we review the facilities provided through R for literate data mining. The benefits in following this process are significant and data miners are encouraged to be literate! We will end up with quality reports, with highly repeatable processes, and all of the code available for others to validate and to follow and build upon.