Universal Revision analysis

[dsweber2]

Initially motivated by a request from Nat on #1439. We have some sources with decent revision behavior discussions, and some without. It would probably be useful to write a script that spits out some basic revision behavior. Unlike some of the other dataset attributes, this one is fairly agnostic to the particulars of the dataset (compare with correlations, for example).

People best equipped to help with this are @lcbrooks and @jingjtang, though I suspect just adapting a script that jinging sent me quite a while back is probably the easiest thing.

In many respects this goes under the general heading of having a more unified structure of dataset analysis. Currently the best place to look for dataset exploration is the issue associated with adding the dataset, which doesn't seem particularly accessible.