Abstract

Computational research increasingly relies on open-source code to enable reproducibility and adoption. Yet little is still known about how code quality, repository engagement, and research impact are associated. This paper investigates two linked research questions focusing on research repositories: (1) whether automated code quality assessments e.g., those measuring readability and computational environment flexibility, correlate with repository engagement metrics (stars and subscribers), and (2) whether these developer metrics are associated with citation counts of associated papers. We analyze papers and research repositories (typically developed by the authors of the papers) from Papers with Code across 2007–2010 - allowing a sufficient time horizon for engagement and impact metrics to have emerged - using an AI-based rubric for systematic code quality assessment across five dimensions with a total …

Date

2026

Authors

Mayank Kejriwal

Book

Proceedings of the 4th ACM Conference on Reproducibility and Replicability

Pages

220-225