Data Mining in the Classroom: Discovering Groups' Strategies at a Multi-tabletop Environment

Large amounts of data are generated while students interact with computer based learning systems. These data can be analysed through data mining techniques to find patterns or train models that can help tutoring systems or teachers to provide better support. Yet, how can we exploit students’ data when they perform small-group face-to-face activities in the classroom? We propose a novel approach that aims to address this by discovering the strategies followed by students working in small-groups at a multi-tabletop classroom. We apply two data mining techniques, sequence and process mining, to analyse the actions that distinguish groups that needed more coaching from the ones that worked more effectively. To validate our approach we analysed data that was automatically collected from a series of authentic university tutorial classes. The contributions of this paper are: i) an approach to mine face-to-face collaboration data unobtrusively captured at a classroom with the use of multi-touch tabletops, and ii) the implementation of sequence mining and process modelling techniques to analyse the strategies followed by groups of students. The results of this research can be used to provide real-time or after-class indicators to students; or to help teachers effectively support group learning in the classroom.