Sentiment Analysis in Jira Software Repositories

Mining Software Repositories is a field of study whose main task is to extract valuable information from a large amount of data available within software repositories. This information about systems and projects can be exploited in different ways improving the development processes in Software Engineering. A new research area makes use of this data for analyzing the software professionals' emotional state and their relationship with different factors such as productivity and quality in tasks. In our study, we applied a supervised classification model to predict sentiments contained in issue comments in open source projects hosted in Jira repositories. The main objective is verifying if sentiments are related to issue comments and factors such as issue resolution types (Resolved or Unresolved) as well as the time of day and the day of the week, in which the comments were written. Our results show that comments in unresolved issues tend to express less positive and more negative sentiments regarding the comments in resolved issues. In addition, we also obtained interesting results regarding the sentiments in comments and the time/day of the week of publication.