Analyzing Scrum Team Impediments Using NLP

In this research, we focus on the impediments encountered by students in capstone projects following the Scrum methodology. Scrummeeting notes were collected in a dataset to permit Scrum roles and instructors to monitor progress and issues. We identified 9 categories of impediments in this dataset: Android, Coding Skills, Debugging, External Factors, Firebase/Database, Git/GitHub, Teamwork, Time Management, and UI/UX Design. We developed a Large Language Model (LLM) to classify these impediments. Natural Language Processing (NLP) has the potential to support software engineering processes. The novelty of this research is that it attempts to identify impediments faced by students' Scrum teams with AI and support students and instructors. The relevance of the approach was discussed with subject matter experts (SME) of the industry. The proposed model is useful in both the academic and industry settings, to identify on-the-fly areas that need attention and, if fixed, would increase team productivity.