Integration of Generative AI in Data Science: A New Frontier in Teaching Efficient Coding Practices

Generative AI has taken the world by storm. How is it possible that AI can write rich essays or produce code snippets that seem so perfect? ChatGPT and generative AI products like YouChat seem like game-changers in many areas, including education, where teaching methods have not changed in decades. We see these new tools as game-changers, as many students can now have a personal assistant to help them learn how to code data science assignments. Generative AI is the new calculator with enormous capabilities to advance human knowledge. Coding is one of the most time-consuming and difficult tasks for any student. Yet by using YouChat, students can instantly generate useful code snippets that are helpful in completing an assignment. Just as reading makes people better writers, in this case, reading sample code snippets generated by YouChat makes students better coders. In this session, we will demonstrate a practical use case of integrating YouChat into a data science course. Students learn how to best use generated code snippets by assessing their correctness and relevance to the task and by modifying the code snippets to help solve a larger problem. We will demonstrate a new platform called CodeBench (https://codebench.cs.rutgers.edu/), developed to integrate generative AI into data science, which is currently being tested with 300+ senior undergraduate students. As data science is interdisciplinary, lessons learned from this study can be applied to many educational areas.