•  
  •  
 

Abstract

This article describes our processes for analyzing and mining the vast records of instructor and student usage data collected by a learning management system (LMS) widely used in higher education, called Canvas. Our data were drawn from over 33,000 courses taught over three years at a mid-sized public Western U.S. university. Our processes were guided by an established data mining framework, called Knowledge Discovery and Data Mining (KDD). In particular, we use the KDD framework in guiding our application of several educational data mining (EDM) methods (prediction, clustering, and data visualization) to model student and instructor Canvas usage data, and to examine the relationship between these models and student learning outcomes. We also describe challenges and lessons learned along the way.

Share

COinS