Friday 7 June 2013

n gram action

I noticed recently that google had added a facility to search their full corpus of scanned literature and archived web pages for specific search terms.
Known in computer science as n grams these are essentially a natural language searches, they're a core part of machine translation systems and have many other important uses.
I'm not going to be as useful. What I am going to do is plot the relative mentions of various wines/regions in the literature corpus so that we can finally answer some of the more pressing vinous questions.

Bordeaux or Burgundy?


Which of the First Growths really deserves the accolade of first amongst equals? Obviously taking 1855 as our starting point.

I think it makes for a rather beautiful little graph. Almost certainly of no real use what so ever, still. Here's the link to their n gram search page, let me know if you come up with any good ones. 

No comments: