GHTRec: A Personalized Service to Recommend GitHub Trending Repositories for Developers

GitHub is one of the largest hosting service platforms for software development, which contains more than 40 million users and 100 million software repositories. GitHub provides a trending page to help software developers discover potential repositories during a period of time. Also, GitHub introduces a feature named “topic” to label repositories. However, GitHub does not explicitly provide user preference information. It is difficult for software developers to find personalized GitHub trending repositories satisfying their preferences. In this paper, we propose a service named GHTRec to recommend personalized GitHub trending repositories for software developers. First, we use a deep-learning method to predict topics for GitHub repositories. Next, we leverage the historical repositories committed by software developers to make recommendation of GitHub trending repositories. Then we evaluate our topic prediction model and recommendation service, and results show that our GHTRec service could recommend trending repositories satisfying developers' topic preferences.