In this paper, the author performs an empirical analysis to study the expressed sentiments in the commit comments from different open source projects and analyze their relationships with different factors such as used programming language, time and day of the week in which the commit was made, team distribution and project approval. To perform this analysis, this study select SentiStrength- a lexical sentiment extraction tool specialized in dealing with short and low-quality texts. In particular, this study analyzed the commit comments of 29 projects those have more than 200 comments. The projects in this dataset were implemented in 14 different programming languages. However, statistical analysis shows that the programming languages except for C, C++, Java, Python, and Ruby do not yield a significant result. Interestingly, this study found that Java projects tend to have a slightly more negative score than projects implemented in other languages. Furthermore, this study investigates the correlation between emotions, day and time of the week when the commit comments were written. The author of this study noticed that 78% of the commit comments were written during the week while only 22% were written during weekends. Moreover, they found that while emotion scores are generally neutral, the emotion of commit comments on Mondays tends to be more negative. To study the relationship between the emotions expressed in commit comments and the geographical distribution of the projects’ developers, they manually grouped the location of all developers associated with the 29 projects into countries and also their corresponding continents. However, they did not find any significant correlation between the emotion score in the commit comments and the number of countries or continents in which each project is distributed.
Contributions of The Paper
Provide useful insight into the emotions expressed in the committed comments of open source projects
Provide actionable insight about the correlation of emotion score in the commit comments with different factors such as programming language, time and day of the week and also the countries and continents in which each project is distributed.
Publisher
MSR'14
Link to The Paper
https://dl.acm.org/doi/10.1145/2597073.2597118
Name of The Authors
Emitza Guzman, David Azócar, Yang Li
Year of Publication
2014
Summary
In this paper, the author performs an empirical analysis to study the expressed sentiments in the commit comments from different open source projects and analyze their relationships with different factors such as used programming language, time and day of the week in which the commit was made, team distribution and project approval. To perform this analysis, this study select SentiStrength- a lexical sentiment extraction tool specialized in dealing with short and low-quality texts. In particular, this study analyzed the commit comments of 29 projects those have more than 200 comments. The projects in this dataset were implemented in 14 different programming languages. However, statistical analysis shows that the programming languages except for C, C++, Java, Python, and Ruby do not yield a significant result. Interestingly, this study found that Java projects tend to have a slightly more negative score than projects implemented in other languages. Furthermore, this study investigates the correlation between emotions, day and time of the week when the commit comments were written. The author of this study noticed that 78% of the commit comments were written during the week while only 22% were written during weekends. Moreover, they found that while emotion scores are generally neutral, the emotion of commit comments on Mondays tends to be more negative. To study the relationship between the emotions expressed in commit comments and the geographical distribution of the projects’ developers, they manually grouped the location of all developers associated with the 29 projects into countries and also their corresponding continents. However, they did not find any significant correlation between the emotion score in the commit comments and the number of countries or continents in which each project is distributed.
Contributions of The Paper
Comments
No response