Applying Learning Analytics in the course TDT4100 at NTNU

Abstract

In the Object Oriented Programming course TD4100 at NTNU there is a need to get an improved overview of the learning challenges of the students. A better understanding of how the students work, and how they handle the weekly programming exercises, could enable further improvements of both the exercises and the course.

A design science approach was used to implement a Learning Analytics system to be used in TDT4100 and similar courses. The aim was to create a visualisation of the students' progress through exercises, by collecting snapshots of the source code being written. Several metrics were extracted from the snapshots to enable an expression for progress to be created. An experiment was conducted with voluntary students in TDT4100 to assess the proposed progress metric.

The results indicated that the progress of students when completing exercises can be measured. The results also indicate that visualising the progress metric over time enables identification of areas of breakdowns. Inspecting the students' source code at these areas was shown to aid in identifying the challenges of the students.

The implemented system worked according to specifications and enabled research into how students progressed through exercises. The system was flexible and several extensions were successfully made through iterations of development and evaluation. The results of the thesis is seen as a starting point for further research into the challenges of the students in TDT4100.