Text mining is one of the ways of doing digital research in the humanities. It allows you to extract key terms or topics and see their continuity or change over time and also to correlate different sources. One helpful tool for doing text mining is Voyant. Voyant is an open source tool that allows you to do a basic text analysis. With it, you are able to extract words from a corpus, see how often they are used, correlate the words with other words, discover themes in the document, etc.
In working with Voyant, you can either ruse their online website or run their server on your computer. It is preferable to download their server and run it in your computer to avoid problems with load speeds due to slow bandwith issues. Downloading and running the server on both the Mac and Windows computers is a breeze. In working with this tool, I uploaded about seventeen files from the Slaves Narrative Collection available through the Library of Congress Born in Slavery project.
In analyzing this collection, Voyant was able to create a word cloud. The word cloud below is created by Voyant based on the files that I uploaded. When you hover the mouse pointer on each of the words, it will tell you how many times the word has been used in the documents.
The documents are I uploaded were based on interviews conducted by the slaves from different states. Voyant is able to graph the number of times that the word was used in each of the states. See the image below:
Voyant can even do more. I wanted to see the relative frequency in the different states when the word “good” was used with “master.” The image below shows that. Though my images do not give the number of times, when you hover the pointer of your mouse on the graph, it would give you the exact number of times.
This tool offers many possibilities for working with large text.