Towards Mining OSS Skills from GitHub Activity

ICSE 2022 |

Open source software (OSS) development relies on diverse skill sets. However, to our knowledge, there are no tools which detect OSS-related skills. In this paper, we present a novel method to detect OSS skills and prototype it in a tool called DISKO. Our approach relies on identifying relevant signals, which are measurable activities or cues associated with a skill. Our tool detects how contributors 1) teach others to be involved in OSS projects, 2) show commitment towards an OSS project, 3) have knowledge in specific programming languages, and 4) are familiar with OSS practices. We then evaluate the tool by administering a survey to 455 OSS contributors. We demonstrate that DISKO yields promising results: it detects the presence of these skills with precision scores between 77% to 97%. We also find that over 54% of participants would display their high-proficiency skills. Our approach can be used to transform existing OSS experiences, such as identifying collaborators, matching mentors to mentees, and assigning project roles. Given the positive results and potential impact of our approach, we outline future research opportunities in interpreting and sharing OSS skills.

[ICSE 22] Towards Mining OSS Skills from GitHub Activity

This is the full-length presentation for the ICSE'22 paper titled "Towards Mining OSS Skills from GitHub Activity" by Jenny T. Liang, Thomas Zimmermann, and Denae Ford. This work is presented by Jenny T. Liang. LEARN MORE: Read our paper: Contact Jenny: Visit Jenny's website: