Text Mining

Below is a Bookworm chart I made with the Chronicling America newspaper collection. I used the search terms “bank,” “credit,” and “currency” for the years 1836-1900. Note the huge spike for the Election of 1840 — followed by precipitous drop in financial issues as a subject of political debate — and then the smaller spikes for Panics in 1857, 1873, and 1893. No big surprises here, but it is nice to see the visualization.


This could be an interesting exercise with students — asking them to search words they think are important, and then see if they were right. (And what do they think it means if they weren’t?) Or they could test different terms and phrases related to their topics, which would improve their ability to work with other primary source databases.

Some “distant reading” on newspapers, court proceedings, business records, etc. related to my Wilmington project would probably be helpful — but most of the relevant sources haven’t been put into appropriate formats. Uncorrected OCR for 19th century newspapers is pretty unreliable.

Leave a Reply

Your email address will not be published. Required fields are marked *