Reproducibility package — Multivariable analysis for biomass transportation cost: a new formulation based on a systematic review of prior cost formulations

Reproducibility package for the manuscript "Multivariable analysis for biomass transportation cost: a new formulation based on a systematic review of prior cost formulations" (Lara-Montaño et al., 2026, submitted to Renewable and Sustainable Energy Reviews). Contents: (i) the PRISMA-2020 internal protocol and decision log; (ii) the raw and merged Scopus and Web of Science exports; (iii) Rayyan screening exports with tiebreaker resolutions and the list of 21 eligible-but-not-retrieved papers; (iv) the structured extraction of 252 papers in 71 fields per paper, including a 20-column binary matrix encoding the structural ingredients of each cost model; (v) the post-audit working CSV after two-pass unit-consistency audit and PPP normalisation to international USD 2026; (vi) the World Bank PPP/FX series used as input for the currency normalisation; (vii) 27 Python scripts implementing the full analysis pipeline (descriptive, density-regime test, HIC vs non-HIC test, Option A reference cells, Option B log-log regression with Block A robustness battery, cost decomposition into fixed + per-km components, physics-based techno-economic reconstruction and scope-ratio regression, Cohen's κ pilot); (viii) a shell driver (reproduce.sh) that assembles the working layout, runs the 17 essential pipeline steps with fixed seeds, and copies regenerated outputs back to the deposit; (ix) the canonical numeric outputs consumed by every claim in the manuscript, persisted as JSON, with a frozen reference snapshot under results_reference/ for byte-exact regression testing; (x) the 11 manuscript figures (Fig. 1-3 plus Fig. 4-11 produced by the pipeline) as vector PDFs, plus exploratory variants. The pipeline is idempotent — random seeds are fixed (seed = 42 for bootstrap and k-fold cross-validation) — and runs end-to-end in approximately 3 minutes on a 2023 Apple M2 laptop.