This analysis allows you to pick two categories of data,
and it produces scatter plots so you can look for correlations
and trends.

First it shows you a separate plot for each of the 24x24 latitude/longitude 
locations (using 'small multiples' on a map).  Each 'small multiple' scatter 
plot shows all of the data available for that location (usually 72 points,
for the 72 months of data), for the selected two categories.  You can click
on the small multiples to see a full-size version of the scatter plot.

Following the small-multiple map, you can scroll down to see an overall
correlation scatter plot, which shows all the data at once.

You can look for correlations in the scatter plots, and also
notice that the areas over water, and the areas over land have
different correlations (and also the areas in mountain areas).
This is easiest to see when you look at the grid side-by-side with a map.

Reference: This idea of having extremely high-density small-multiple
plots all on the same page was somewhat derived by a sample that
Jennifer Sloan had created.  (not the code, but the idea :)