Using Stratified Attribute Tracking (SAT) Diagrams for Learning Analytics

We have created a visual representation called Stratified Attribute Tracking (SAT) Diagram to explicate trends that are otherwise implicit in learning analytics data. SAT Diagram is a unified graph that enables tracking individual attribute values in a dataset and stratifying them according to criteria set by the researcher. SAT diagram represents the transition of samples between strata across attributes. In this paper we introduce the SAT diagram and illustrate how to generate, interpret and analyze them. We believe the process of SAT diagram generation would enable exploring deeper research questions on learning data.