Correlation of User Behaviour Patterns and Assignment Supplements in KTH-GitHub Repositories

With the increased popularity and usage of GitHub, universities such as KTH have chosen to include it in their Computer Science curriculum. Furthermore, the interest for mining GitHub-repositories has shared a similar increase. The purpose of this study is to investigate if commit metadata is suffic...

Full description

Bibliographic Details
Main Authors: Isacson, Jonas, Lindblom, Emil
Format: Others
Language:English
Published: KTH, Skolan för datavetenskap och kommunikation (CSC) 2017
Subjects:
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-208558
Description
Summary:With the increased popularity and usage of GitHub, universities such as KTH have chosen to include it in their Computer Science curriculum. Furthermore, the interest for mining GitHub-repositories has shared a similar increase. The purpose of this study is to investigate if commit metadata is sufficient enough to perform a distinct categorization of KTH students’ GitHub user behaviours for course assignments. Furthermore, the study investigates if there is a correlation between the various categories and the number of assignment supplements. The KTH GitHub organization was mined after which the average user behaviour was calculated for each assignment. Students were then compared to the average usage behaviour and categorized based on 5 different user behaviour segments. Moreover the average number of supplements was calculated for each category. The results show that there is no evident correlation between user behaviour patterns and assignment supplements. However, comparisons between course rounds reveal similar tendencies in represented categories, their average supplement count and their populations.  === Med den ökade populariteten och användandet av GitHub har universitet som KTH valt att inkludera det i deras läroplan. Intresset för datautvinning av GitHub-repositories har även sett en ökning på senare år. Målet med denna studie är att undersöka om commitmetadatan är tillräcklig för att kategorisera KTH studenters användarbeteende under kursuppgifter. Studien undersöker även om det finns en korrelation mellan kategorierna och antalet uppgiftskompletteringar. Data utvanns ur KTHs GitHub-organisation varpå det genomsnittliga användarbeteendet beräknades för varje uppgift. Studenterna jämfördes sedan med den genomsnittliga användarbeteendet och kategoriserades baserat på 5 olika användarbeteendeområden. Det genomsnittliga antalet kompletteringar beräknades sedan för varje kategori. Resultatet visar på att det inte finns en tydlig korrelation mellan användarbeteende och uppgiftskompletteringar. Jämförelse mellan kursomgångar visar emellertid liknande tendenser i representerade kategorier, deras genomsnittliga kompletteringsmängd samt deras population.