Monitoring Student Activity in Collaborative Software Development

TL;DR

This study analyzes student activity data from a collaborative software development course to identify metrics and techniques for proactively detecting low engagement and improving peer collaboration.

Contribution

It introduces a data-driven approach to monitor and analyze student engagement and collaboration patterns using multiple software tools and interaction traces.

Findings

Identification of key engagement metrics

Preliminary analysis of collaboration patterns

Potential for proactive intervention

Abstract

This paper presents data analysis from a course on Software Engineering in an effort to identify metrics and techniques that would allow instructor to act proactively and identify patterns of low engagement and inefficient peer collaboration. Over the last two terms, 106 students in their second year of studies formed 20 groups and worked collaboratively to develop video games. Throughout the lab, students have to use a variety of tools for managing and developing their projects, such as software version control, static analysis tools, wikis, mailing lists, etc. The students are also supported by weekly meetings with teaching assistants and instructors regarding group progress, code quality, and management issues. Through these meetings and their interactions with the software tools, students leave a detailed trace of data related to their individual engagement and their collaboration behavior in their groups. The paper provides discussion on the different source of data that can be monitored, and present preliminary results on how these data can be used to analyze students' activity.