Replication Package for "Modeling OSS Popularity and Survival: A Non-Linear and Structural Analysis of GitHub Projects"

This replication package accompanies the manuscript titled “Modeling OSS Popularity and Survival: A Non-Linear and Structural Analysis of GitHub Projects”. It includes curated datasets, preprocessing scripts, and modeling pipelines used to investigate the non-linear dynamics of OSS popularity and the structural predictors of project survival. The dataset comprises over 60,000 GitHub repositories sampled across multiple time windows (3m, 6m, 9m, 12m), with features spanning development activity, engagement metrics, readability proxies, and documentation presence. All experiments were conducted using Python 3.11 with reproducible scripts and fixed random seeds. The name field has been removed to preserve project neutrality and avoid reputational bias. This package enables full replication of the GAM and Weibull AFT models presented in the manuscript.