Sentiment Analysis in monitoring software development processes: An exploratory case study on GitHub's project issues

Abstract

Software process models, which allow us to develop software products, can be improved by using the corresponding quality model. However, current tendencies in the application of Global Software Engineering and Global Software Development, which forces geographically dispersed teams to collaborate, make the usual monitoring techniques obsolete. This situation has led to looking for new methods that can help in the decision making process, such as the case of the Social Network Analysis field.In this article we propose the introduction of Sentiment Analysis techniques in order to identify and monitor the underlying sentiments in the text written by developers in issues and tickets. Therefore, in order to check its viability we conducted an exploratory case study analysing polarity and emotional clues in development issues from nine well-known projects that are freely available. Results show that although both polarity and emotional analysis are applicable, the emotional analysis looks to be more suitable to this kind of corpus. The developers leave underlying sentiments in the text, and that information could be monitored as any other feature in the development process.