Profile of Open Source Project Contributors
Scientific research — recognized as the best paper of our 6th-period interdisciplinary work at PUC Minas. Analyzing 1,000 repositories from GitHub's most popular languages with the Goal-Question-Metric (GQM) paradigm, the study found contributors concentrated in the United States, India, and China (Brazil eighth), mostly at an intermediate experience level (88.41%), with experienced contributors reaching an 83.76% average pull-request approval rate.
This was TI6, the sixth-period interdisciplinary project in the Software Engineering bachelor's at PUC Minas. Instead of shipping another product, the brief asked us to do research: pick a question that mattered to the field, gather evidence, and write it as a scientific paper.
I co-authored the work with Arthur F. Costa, Fernando A. F. Ibrahim, Gabriel P. Carvalho, Luca F. Azalim, and Renato M. A. Penna. The paper follows the structure of empirical software-engineering writing — related work, Goal-Question-Metric design, results, threats to validity, conclusion — in the SBC format we were trained to use.
The writing itself was the project. The paper was recognized as the best of the period: not for a demo, but for how clearly we argued from data.
We mined the 100 most-starred repositories in each of GitHub's ten most popular languages in 2024, then looked at 52,743 contributors over twelve months. The Goal-Question-Metric paradigm turned a broad curiosity — who actually contributes to OSS, and does experience change how they participate? — into five research questions on geography, affiliation (including MANGA companies), experience, pull-request approval, and contribution frequency.
The artifact below is the full paper. It is written in Portuguese, with an English abstract.