Sarah Payne, William Quinn, and Avery Blankenship Our panel demonstrates various applications for text mining and literature, with a particular focus on how text mining tools illuminate our understandings of race, gender, and class in 19th and 20th century literature. The first paper will address the use of RStudio to perform a sentiment analysis of 20th century passing novels. Using both sentiment analysis and RStudio visualization tools, this paper will examine whether it is possible to track and visualize a conception of race that goes beyond, and perhaps refuses, the black/white racial binary. The second paper will use both BookNLP and Python to try and create a computational version of the Bechdel test. Applying this Bechdel test to nineteenth century novels demonstrates the risk of encoded gender bias, as machine learning methodologies still rely on training data provided by human programmers. Ultimately, this paper argues for more gender diversity in digital humanities and computational fields. The final paper focuses on Ezra Pound’s departure from the editorial board of Poetry, a Magazine of Verse, and his accusation that the magazine had become a “meal-ticket” for mediocre poets. Contrary to Pound and others’ vision of the isolated poetic genius, an idea entangled in class, computational tools can recuperate the collective collaboration within magazines. Using doc2vec, this paper shows how a communal form of production, based on literary recycling and reuse, was a central aspect of modernist poetry. While we welcome those experienced in digital methods, we aim to make our panel accessible to those with little to no text mining experience. Our panel will provide a user-friendly approach that balances methods with interpretation. We hope to introduce both students and practitioners to possible uses of text mining and demonstrate the benefit, and even necessity, of applying humanistic methods to digital technology.
