It’s not the easiest thing to search for on Google. But that is not stopping companies (like Pure Visibility!) from taking advantage of the open source analysis software R for client and internal analytics, from statistics to machine learning. So why aren’t you using it?

Or maybe you are. While proprietary statistical packages like SAS once ruled the corporate world, a recent New York Times article suggests that not everyone fears what is free (R is the freeware version of S+). Google uses it for analysis of search patterns, Pfizer for scientific research.

R can do everything from basic statistical testing for significance, ANOVA, etc, to more sophisticated regression, to other data mining methods including constructing Bayes Networks, or Decision Trees. The uses for web analysis, then, are too many to count!

Here’s a quick summary table of Needs, the R features that can be used to address them, and the requirements:

Need	R Features & Methods	Requirements
Basic statistical testing, analysis of variance, descriptive stats	ttest(), summary(), aov(), TukeyHSD()	Understanding of when to use each test
Regression	glm(), mle(), maxLik()	Knowledge of performing basic line-fitting – modeling error, etc
Exploratory Data Analysis	spatstat package, bayes nets package, and others	Knowledge of exploratory data mining, such as how to set a window size, etc
Documentation on anything R	rseek, the R search engine	Patience – sometimes you have to surf to find the functions and examples you need!
Other packages, or to download R	CPAN, the source for all things R

Are you using R? Let us know how!