Mining learners’ traces from an online collaboration tool

As university courses increasingly require students to use online tools in their studies, the opportunity arises to mine the resulting large amounts of student learning data for hidden useful information. In this paper we study the application of data mining to data collected from third year software development group projects using Trac, an online collaboration tool. We applied two very distinctive techniques, clustering and sequential pattern mining. The results point to the importance of leadership and group interaction, and give promising indications of whether effective leadership is occurring in a group. In addition, patterns were found which appear to indicate good individual practices. The results have considerable promise for advising groups at the start of their work and in early identification of both effective and poor patterns, in time for remediation.