Understanding Open Source Contributor Profiles in Popular Machine Learning Libraries
Jiawen Liu, Haoxiang Zhang, Ying Zou
TL;DR
The paper tackles how to characterize contributors to open-source ML libraries by deriving activity-based profiles from large GitHub datasets. Using 7,640 human contributors across six ML libraries, it identifies four profiles (Core-Afterhour, Core-Workhour, Peripheral-Afterhour, Peripheral-Workhour) via clustering on a compact feature set, and then analyzes OSS engagement through workload patterns, work preferences, and technical importance. It further links contributor engagement to project popularity through mixed-effects models, finding that balanced contributions and robust code-review activity correlate with higher stars and forks, while early-stage dynamics and core-periphery distinctions modulate these effects. Longitudinal analysis reveals that some long-term contributors evolve toward fewer, constant, balanced, and less technical contributions, suggesting pathways for onboarding and retention. Overall, the work provides a data-driven framework for understanding ML OSS contributors, with practical implications for project management, contributor retention, and future research in software engineering for ML ecosystems.
Abstract
With the increasing popularity of machine learning (ML), many open-source software (OSS) contributors are attracted to developing and adopting ML approaches. Comprehensive understanding of ML contributors is crucial for successful ML OSS development and maintenance. Without such knowledge, there is a risk of inefficient resource allocation and hindered collaboration in ML OSS projects. Existing research focuses on understanding the difficulties and challenges perceived by ML contributors by user surveys. There is a lack of understanding of ML contributors based on their activities tracked from software repositories. In this paper, we aim to understand ML contributors by identifying contributor profiles in ML libraries. We further study contributors' OSS engagement from three aspects: workload composition, work preferences, and technical importance. By investigating 7,640 contributors from 6 popular ML libraries (TensorFlow, PyTorch, Keras, MXNet, Theano, and ONNX), we identify four contributor profiles: Core-Afterhour, Core-Workhour, Peripheral-Afterhour, and Peripheral-Workhour. We find that: 1) project experience, authored files, collaborations, and geographical location are significant features of all profiles; 2) contributors in Core profiles exhibit significantly different OSS engagement compared to Peripheral profiles; 3) contributors' work preferences and workload compositions significantly impact project popularity; 4) long-term contributors evolve towards making fewer, constant, balanced and less technical contributions.