Finding behavioral indicators from contextualized commits in software engineering courses with process mining - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Finding behavioral indicators from contextualized commits in software engineering courses with process mining

Résumé

Git4School is a dashboard helping teachers to monitor and make decisions during Git-based lab sessions in higher education computer science programs. This tool makes it possible to visualize the commits made by students over time according to the context and, in particular, the type of pedagogical intervention by the teacher (discussions between students on the problem, dissemination of a solution, etc.). Despite its visualizations providing indicators for decision-making, the tool does not provide information about the student's behavior. There are existing studies dealing with Process Mining (PM) in education, specifically in computer science courses and using Git. Through an empirical exploratory study, we explore the possibility of taking advantage of these contextualized commits using PM. We analyzed data from 5 teaching units covering different higher education levels using the bupaR library. Firstly, we discovered promising indicators to predict students' behavior during a lab session. Secondly, we identified several possibilities for future research on PM and contextualized commits. Finally, we have established a set of recommendations to help analyze contextualized commits using PM.
Fichier principal
Vignette du fichier
paper.pdf (490.07 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04332205 , version 1 (08-12-2023)

Identifiants

Citer

Mika Pons, Jean-Michel Bruel, Jean-Baptiste Raclet, Franck Silvestre. Finding behavioral indicators from contextualized commits in software engineering courses with process mining. 2nd International Workshop on Frontiers in Software Engineering Education (FISEE 2023), Jan 2023, Villebrumier, France. pp.56-68, ⟨10.1007/978-3-031-48639-5_5⟩. ⟨hal-04332205⟩
107 Consultations
24 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More