Chester Ismay (Portland State University), Arturo Valdivia (Indiana University)
Abstract
This workshop presents applications of Large Language Models (LLMs) in data courses, highlighting how generative AI tools can spark student and instructor curiosity and deepen understanding while reducing instructor workload. Using examples from ModernDive (https://moderndive.com/v2), participants will explore strategies to develop engaging lesson plans around core statistical concepts while integrating LLM-driven techniques that enhance both teaching and learning. By automating routine tasks, generating starter code and relevant images, and offering real-time feedback, LLMs have the potential to reduce the time instructors spend on repetitive chores.
By walking through real-world examples, attendees will learn how to leverage GitHub Copilot in RStudio to generate, refine, and explain code in tandem with traditional statistical instruction, helping students write code quickly and efficiently. We will demonstrate how to blend the power of LLMs with data workflows, empowering students to move from conceptually grasping to hands-on data exploring. We will also illustrate how AI-generated suggestions can be turned into class discussions and how to generate multiple-choice questions to check student understanding. Finally, we will reflect on the benefits and limitations of these emerging tools.
By the session’s end, participants will have concrete ideas for course design that pair foundational statistical and learning sciences material with cutting-edge LLM assistance. This dual approach enriches the learning experience—stimulating student curiosity, building coding proficiency, and fostering problem-solving creativity at the heart of statistics. Ultimately, this workshop equips educators with a forward-thinking toolkit for both themselves and their students by encouraging curiosity.