Skip to Digital Humanities site navigationSkip to main content

Topic Modeling the Jane Addams Digital Edition (2024)

by Olivia Church (Math Major)

Four people in front of a projected slide "Exploring Jane Addams Documents through Topic Modeling and Multilable ClassificationOlivia Church used data from documents in the Jane Addams Digital Edition as the basis for her exploration of topic modeling. This is a computational process by which Olivia counted and analyzed the words in Addams’s speeches and articles to extract topics, or themes. Words that appear together often are clustered into topics that are then evaluated by subject experts to add meaning. Once documents have topics they are further analyzed using other data fields to indicate when topics were more widely used, whether in speeches or articles, and such. The “peace” topic appeared most frequently in Addams’s writings (in 52.6% of documents). One research question she addressed was to see whether human analysis and computer analysis generated similar results, and while there was not a 100% overlap, the signs were that topic modeling could be used to successfully predict the subject that would be assigned to a document.