Formal in the Informal: A Multi-Level Analysis of Core Python Developers' Tweets

Abstract

Context: Contributors to Open Source Software (OSS) development projects use social media to discuss personal and work-related issues. While there has been some interest in studying social media posts of developers, the investigation of the use of this medium by core contributors has not received much attention. Such an investigation could be useful given the technical and social insights such members bring to their team's development. Objective: To bridge this gap, we have investigated Python core developers' posts on Twitter. Method: We systematically collected 27,805 tweets from 80 core developers' Twitter accounts of which 3,282 (11.8%) were relevant to Python development. We performed multiple forms of analyses on these tweets, involving bottom-up thematic analysis on three levels (topics, questions and answers, and decision-making). We also conducted social network analysis of the topics of interest of core developers. Results: Core developers' Python-related tweets fall under 23 different topics (in 5 higher-level topic categories). The top two categories were software-related and community-related tweets. Core developers also used Twitter as a medium to discuss technical issues related to Python software development in the form of questions and answers (Q&As) to a large extent (45% of the tweets). Moreover, this informal medium is also used for decision-making 2% of the time, supplementing the formal channels of decision-making (i.e., mailing lists). Additionally, we have identified five different groups of core developers based on their topic interests. Conclusion: Although informal in nature, social media platforms provide another window, beyond formal project repositories, for understanding OSS ecosystems. Consequently, informal sources must be integrated with formal sources to obtain a comprehensive view of OSS development activities.