Enriched GitHub Repositories Dataset: Popularity, Structure, and Language Diversity (2011–2020)

This dataset is an enriched and aggregated compilation of GitHub repositories created from queries to public databases using BigQuery, specifically bigquery-public-data.github_repos and githubarchive. Rather than using the raw tables directly, the data has been combined and processed to build a more informative resource.

It contains information about relevant GitHub repositories, including their popularity, internal structure, content size, proportion of binary files, programming languages used and the licence applied. In addition, I included several derived metrics, such as language entropy, the proportion of the dominant language, average path depth, and the first year in which activity was recorded in the GitHub Archive.

The main goal is to support the exploration of relationships between different aspects of open-source projects. In particular, it is designed to analyse how popularity relates to structural complexity (e.g. number of files, content size or project depth), how repositories differ in terms of programming language diversity and specialisation, and how these characteristics vary depending on the age of the project.

The final table contains 5,772 repositories and combines mainly categorical and numerical variables.