I spotted this on Luke McKernan’s excellent blog which deals with silent and early cinema which can be found at https://bioscopic.wordpress.com/. I can do no better than take Luke's words: A new tool from Google Labs offers interesting ways of analysing silent film subjects (or any other subject, for that matter). The Ngram Viewer uses data taken from the 15 million books and other documents scanned by Google Books to trace the occurrence of words or phrases (up to five words) between 1800 and 2000, showing how often they occur each year.
All you do is enter your search term or phrase, then choose a time period and your language, and you get the results presented as a graph. Having searched for your term, below the graph you are given the option to search Google Books itself for your term by particular time periods or universally. You can also compare your term with others, by adding a comma-separated second term into the search box. You can compare any number of terms, though there are only five colours available.
There seem to be any number of interesting applications for this as a tool, even if the results are approximate and erratic. The frequency of appearance of terms in books is not necessarily a reliable guide to their importance, and some terms register no scores at all (e.g. Gaumont, Muybridge, Mary Pickford), presumably because Google Books hasn’t indexed them yet. But there is more than enough there to encourage imaginative searches and to yield interesting discoveries.
I have included some examples here. Most of us have some awareness of the origins of the word Photography and a Ngram produces an expected graph. If one adds ‘imaging’ then it’s clear how, with the rise of digital photography, this word has begun to supersede photography. Another interesting graph was produced by showing platinotype and palladiotype (the subject of discussions on BPH) the others are self-explanatory.
Although there are issues with the data (it’s dependent on the books google has scanned and the use of which words or phrases to use needs careful consideration) this could be a useful tool for representing trends
Do have a go, and let me know of any interesting Ngrams that you are able to create.
