Freelance Unbound

Jump to Comments

Tired of the short-term focus of Twitter’s #trending topics? Here’s a great geeky toy to play with over Christmas – Google Books’ Ngram Viewer.

The Ngram Viewer allows you to enter search terms and see how frequently they are used over time. This is a kind of historic version of Twitter trends – nothing after 2008 at the moment, but its results can stretch all the way back to 1700.

What can we see? How about “engagement” as a search term?

Clearly “engagement” was at its peak in the decade straddling the year 1780. Then it underwent a more-or-less two century decline until 1980, before mysteriously spiking up again. Why? It became much less important socially that people were getting engaged to be married – but much more important that they enjoyed whatever it was they were doing.

Another interesting example is “digital”, using just the American English dataset. You’d rightly expect digital as a reference in books to come into its own with the advent of digital computing – but what are these small spikes in usage doing in the late 18th century?

It seems this was a fertile time for medical discovery – with herbalists and anatomists investigating the medical value of digitalis and how your fingers work.

Google is generating datasets for all the books it scans – from single word usage up to multi-word phrases. At the moment, single word datasets are the most complete in all languages Google is working with. Double-word datasets and above are in progress.

If you want to really alienate your family this Christmas, you can download the Google datasets here and start running your own experiments…

Tweet