HFCommunity: A Tool to Analyze the Hugging Face Hub Community

In recent years, empirical studies on software engineering practices have primarily relied on general-purpose social coding platforms such as GITHUB or GITLAB. With the emergence of Machine Learning (ML), platforms specifically designed for hosting and developing ML-based projects have appeared, being HUGGING FACE HUB one of the most popular ones. HUGGING FACE HUB focuses on facilitating the sharing of datasets, pre-trained ML models and applications built with them (spaces in HUGGING FACE HUB terminology). Besides, the Hub is adding more and more collaborative features, such as issues and pull requests, to facilitate the building of these artifacts within the platform itself. With over 100K repositories, and growing fast, HUGGING FACE HUB is therefore becoming a promising source of data on all aspects of ML projects and the community interactions around them. As such, we believe it is a promising source for all types of empirical studies aimed at analyzing the collaborative development and evolution of ML artifacts. Nevertheless, apart from the API provided by the platform, there are no easy-to-use solutions to collect and explore the different facets of HUGGING FACE HUB data, including the repositories, discussions and code evolution. To overcome this situation, in this paper we present HFCOMMUNITY, a relational database populated with HUGGING FACE HUB data to facilitate empirical analysis on the growing number of ML-related development projects.