What is the Problem?

There is no shortage of large, do-it-all statistics programs that contain all the statistical tests you need to do your associations and correlations (and 10,000+ other statistical tests all rolled up into one huge package).

The problem with these is that you have to run the analyses manually. OK, sure, it's not terribly difficult to run a two-sample t-test or an ANOVA – they're just a few clicks away – but let's have a look at this in the context of the whole dataset.

Let's take a dataset of 100 variables (columns). That's pretty small by today's standards. The number of possible relationships in this dataset is (100-1) + (100-2) + (100-3) + … + (100-98) + (100-99) = 4,950.

Now let's suppose that it takes an average of just five mouse clicks to locate and select your chosen test, select your variables, and click GO. That means that the story of your data is 4,950 x 5 = 24,750 mouse clicks away. Gulp.

But what if you...