| K. Stamou, C. Sgouropoulou |
| Explainable Collaborative Analytics: Empowering Educators to Evaluate and Intervene in Digital Collaboration Scenarios |
| 11th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM 2026), Thessaloniki, Greece, September 18-20, 2026 |
|
ABSTRACT
|
| There is a strong move for Learning Analytics dashboards to represent the state of collaborating student groups using just one number, one color or one ranking. When these are used by a teacher in the classroom in real-time to make instructional decisions, the models that produce them are opaque: a teacher knows a group is struggling, but not why. This paper proposes Explainable Collaborative Analytics (XCA), a framework that blends collaboration analytics and Explainable Artificial Intelligence (XAI) to ensure group-level indicators are always paired with explainable evidence, that teachers can examine, question, and react to. XCA has three layers of explanation, namely feature-level, interaction-level and pedagogical-level, as well as a catalogue of intervention templates to convert an explanation into a teacher action. We explain the design principles, architecture and how explanations are tied to the collaboration environment¢s signals. Using a scenario in a project-based classroom, we illustrate how a poor performing group indicator is elaborated with explanation about role-imbalance and lack of argument-driven discussion evidence which supports teachers in targeted intervention. XCA provides the means to move collaboration analytics from opaque scoring towards a meaningful, teacher-centered platform for evaluation and intervention.
|
| 18 September, 2026 |
| K. Stamou, C. Sgouropoulou, "Explainable Collaborative Analytics: Empowering Educators to Evaluate and Intervene in Digital Collaboration Scenarios", 11th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM 2026), Thessaloniki, Greece, September 18-20, 2026 |
[
BibTex] [
Print] [
 Back] |
