Sentiment analysis over collaborative relationships in open source software projects

• Main Authors: LI, Lingjia, CAO, Jian, LO, David
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2020
• Subjects: Collaborative and social computing; GitHub; Human factors; Project management; Sentiment analysis; Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/5929; https://ink.library.smu.edu.sg/context/sis_research/article/6932/viewcontent/SentimentAnalysis_2020_30.pdf
• Institution: Singapore Management University

Sentiments and collaboration efficiency are key factors in the success of the open source software (OSS) development process. However, in the software engineering domain, no studies have been conducted to analyze the effect between collaborators' sentiments, and the role of sentiment in collaborative relationships during the development process. In this study, we apply sentiment analysis and statistical analysis on collaboration artifacts over five projects on GitHub. We use sentiment consistency to quantify the relation between sentiments in collaborative relationships. It is found that sentiment consistency is positively correlated with the closeness of collaborative relationships and collaborators' overall sentiment states. We also perform the Granger causality test and network analysis to study the impact of sentiment consistency on a time series basis. It is found that positive consistent sentiments not only improve collaboration willingness to the utmost extent, followed by inconsistent and negative consistent sentiments, they also boost the closeness of the entire project community. These findings can be applied to develop better OSS project monitoring tools and improve project management by taking developers' sentiments during collaborations into consideration.