I've used the ngram stats package to compare the words in a range of documents. I used the rank module to compare ngrams ranking from different documents rather than comparing different rankings from the same document.
I'd welcome any feedback, especially if anyone thinks that I've applied the tools/ interpreted the results incorrectly!
Great project - I'd encourage you to also mail the NSP mailing list to get comments (if you are still interested). http://tech.groups.yahoo.com/group/ngram/