Text as data: Using Redivis and AI to analyze historical news sentiment

This hands-on workshop introduces social scientists to Redivis as a platform for text-based research, using an AI-powered sentiment analysis task as a working example. Participants will use Redivis's no-code transform interface to prepare a real dataset of historical newspaper coverage using a subset of the Library of Congress's Chronicling America collection, then run a guided Python notebook that applies a modern language model to measure shifts in tone over time. This is a practical, applied session focused on building and running a real analysis pipeline. No prior programming experience is required, though some familiarity with Python will be helpful.

Speaker Adelynn Shirts is a member of the Redivis team, where she works directly with researchers and organizations to support their use of data for research and analysis. Her work spans research support, data workflows, technical documentation and education resources, with a focus on helping researchers use Redivis effectively for data preparation, analysis and reproducible research.

She holds both a a bachelor’s degree in Sociology and master’s degree in Data Analytics from Utah State University, with her master’s studies through the Department of Mathematics and Statistics focused on statistics, machine learning, and data science.