Results for "Environmental Impact of CI/CD Pipelines: A Scenario-Based Analysis of GitHub Actions"

This archive contains the collected and derived outputs used to replicate the analyses in "Environmental Impact of CI/CD Pipelines: A Scenario-Based Analysis of GitHub Actions".The dataset covers the paper’s 2024 study window (2024-01-01T00:00:00–2025-01-01T00:00:00). After extracting the archive at the repository root, paths match those expected by the analysis scripts (e.g. results/2024/2024-01-01T00:00:00-2025-01-01T00:00:00/). Contents results/ — Primary pipeline outputs for the study window: • gha_sample.json — sampling metadata for estimating total GitHub repositories, public repositories, and repositories using GitHub Actions • gha_repos/<REPO_NAME>/ — per-repository crawl data (info.json, workflows_runs.json) for ~19,800 sampled repositories • carbon_footprint_<SETTING>_<REGION>.json — estimated carbon and water footprints under three extrapolation scenarios (Baseline, Stratified, Inclusion_Reweighted) across regional grid settings • resource_usage/<REPO_NAME>.json — telemetry from reproduced workflow runs (CPU, memory, network, execution time) • reproducible_2024.json — list of repositories treated as reproducible for instrumentation • execution_time.json, totals.json — job-level and aggregate execution-time summaries used for footprint modeling and downstream statistics stats/stats_results/ — Cached and exported analysis intermediates: • execution_time_stats.json — job, step, and YAML-match statistics • repo_size_cache.pkl — repository clone sizes used for action statistics and resource-usage modeling • test_tool_commands_*.txt — exported test-related shell commands (e.g. pytest, make test, npm test) stats/plots/ — Generated figures from the stats analysis suite, including regional carbon/water plots (Baseline, Stratified, Inclusion_Reweighted) and sensitivity-analysis PDFs under stats/plots/sensitivity/ resource_usage/plots/ — Figures from resource_usage/get_stats.py, including execution-time distributions, resource-usage regressions, and sample-inclusion (IPW) diagnostics stats/.workflow_sanitize_cache/ — Cached workflow-sanitization intermediates used during analysis