Digital Humanities teaches text analysis with Voyant Tools

By Sarah Daponde Staff Writer The Center for Digital Humanities held an event in-person and over Zoom on Nov. 1 to teach attendees about Voyant Tools, a text analysis software. The event was hosted by Bartholomew Brinkman, English professor and the director of the Center for Digital Humanities. He was assisted by Digital Humanities Intern Spencer Pearson. Brinkman began the event by asking who had used Voyant Tools before and the attendees said it was new to them. To demonstrate how the software works, Brinkman selected a story he thought everyone would be familiar with. He chose Mary Shelley’s “Frankenstein,” a text that he preemptively loaded in Voyant Tools. He projected his screen to show how Voyant Tools constructed several different types of text analysis on the novel. This analysis included information on readability, average words per minute, and a word cloud that showed the most popular words in the corpus. The word “corpus” has the same root as the word “corpse” and means “the body of the text,” Brinkman said. The first tool Brinkman showed was a word cloud - a visual displaying the words used in the text, arranged by size according to frequency. “Those things are related to things like vocabulary, density, readability, average words, percentage, and things like that,” he added. To see more context of when the words appear in the corpus, Brinkman showed the other tools, which work together to give a deeper analysis. “A lot of these tools talk to each other,” added Brinkman. “It's generating the kinds of questions we want to follow up on,” said Brinkman. “And it’s generating the kinds of questions that maybe we wouldn’t have asked otherwise.” Pearson said he found Voyant Tools useful for analyzing his own essays. “I just threw in an essay that I had written at some point and was basically just able to see the general gist of what it was about,” he said. Brinkman added that he frequently uses Voyant Tools in some of the courses he teaches when looking at difficult readings. He said he encourages his students to use the software to see the “big picture” of the text right away. “What are the words that jump out at you?” he said. “And then those are potentially themes and things like that, right?” Brinkman also encourages his students to put their own essay drafts into Voyant Tools to see what topics in the paper are “jumping out” the most. He added that this was a helpful way to see what topics need to be written about in greater detail to strengthen the essay. “There are actually quite a few tools,” said Brinkman. “But we’re not going to talk about all of them today.” He added that some tools are more “polished” than others, but that they can all be helpful in their own way for searching for connections within the corpus of a text. One of the attendees, Halcyon Mancuso, executive director of the Mancuso Humanities Workforce Preparation Center, asked Brinkman if he told his students to use Voyant Tools before or after they had read a certain text. Brinkman said he had done it both ways, depending on what he wanted his students to pay attention to, but he was beginning to favor having the students put the text into Voyant before they read. “With the exception of, you know, great students like Spencer here, sometimes it’s difficult to get that initial read right,” he said. Brinkman said it can help when trying to understand the “big picture” of texts such as argumentative essays and nonfiction work, so students could ask more “probing questions.” “I think maybe I used to think that was cheating more than I do now,” he added. He related this to how he is more and more frequently asking his students to use ChatGPT to understand texts. Pearson added he had only used Voyant Tools in his Digital Humanities class so far, but that some of his teachers have him use ChatGPT for the same kind of analysis. Brinkman said he encourages students to use Voyant in any class it would be useful.