Modern educational environments increasingly generate digital traces of learning through platforms for continuous formative assessment, such as online exercises, micro-quizzes, and interactive learning tools. This information thus allows instructors to observe how students’ knowledge evolves during a course.

The workshop introduces participants to techniques on how to interpret such signals as measurements of a dynamical system, where the internal state represents student knowledge and engagement. This perspective opens the possibility of performing Learning analytics, and thus (using control-lingo) applying system identification, modeling, and feedback design methods from control theory to better understand and improve teaching and learning processes.

The workshop combines thus short lectures with demonstrations and hands-on activities using example datasets and open-source tools, and will push the participants to explore how concepts from control and system identification can be applied to learning analytics and educational data.

Participants will learn how to:

• Use platforms for continuous formative assessment to collect classroom data
• Extract and preprocess educational data from learning platforms
• Apply basic system identification techniques to model learning dynamics
• Explore advanced modeling approaches such as state-space models and knowledge graphs
• Discuss how feedback and control concepts can inform teaching interventions

This workshop is intended for:

• Researchers in control theory and system identification
• Engineers and data scientists interested in learning analytics
• Educators interested in data-driven teaching methods
• PhD students and early-career researchers exploring new application domains

No prior experience in educational data analysis is required.

start time – start time + 00:15
Introduction: The classroom as a dynamical system

start time + 00:15 – start time + 01:00
Collecting and preprocessing classroom data

start time + 01:00 – start time + 01:45
System identification of learning dynamics

Break

start time + 02:05 – start time + 02:45
Advanced models for knowledge dynamics

start time + 02:45 – start time + 03:20
Feedback design for teaching interventions

start time + 03:20 – start time + 03:30
Discussion and wrap-up

Damiano Varagnolo
Norwegian University of Science and Technology (NTNU)
University of Padova

This repository contains the material used during the workshop:

• Example datasets from continuous formative assessment systems
• Python notebooks demonstrating modeling techniques
• Slides used during the lectures

Participants are encouraged to bring a laptop in order to experiment with the notebooks.