Embodied Agents to Scaffold Data Science Education.

Arguing and working with data has become commonplace in several study domains. One way to immerse students in hands-on exploration with data is to provide them with problem-solving environments, for example jupyter notebooks, which can scaffold students' reasoning and bring them closer to disciplinary ways of thinking. Although the intrinsic affordances of jupyter notebooks (e.g., interaction with multiple data representations, automation of procedural task aspects) allowstudents to engage in rich learning experiences, students lack crucial social scaffolding that directly targets the process of learning. We are developing an AIED infrastructure EASEx for use in higher education contexts that brings in the affordances of embodied pedagogical agents to significantly advance educational practice by scaffolding students in a personalized manner as they work through problems using jupyter notebooks.