Replication package: Eco-efficiency and decoupling in data-sparse resource regions — a measurement protocol illustrated for Saudi Arabia's Eastern Province

Data and code to reproduce every quantitative result of the paper "Eco-efficiency and decoupling in data-sparse resource regions: a measurement protocol illustrated for Saudi Arabia's Eastern Province." A single Python script (replicate_all.py) regenerates all tables, all figures, and all perturbation experiments from one input file: the indicator efficiencies, the dispersion-based weights under both the coefficient-of-variation and entropy schemes and under both orderings of the weighting and normalization steps, the composite trajectory across the eco-efficiency plane, the decoupling classification under both the six-state and the original eight-state Tapio partitions, and three sensitivity analyses — perturbation of the allocation key, exclusion of sample boundaries, and joint propagation of uncertainty in the output series, the allocation key, and the weighting scheme. Random seeds and draw counts are fixed, so the Monte Carlo results are reproducible to the digit. The dataset covers the Eastern Province of Saudi Arabia over 2015–2023. Gross value added is aggregated to annual frequency from the quarterly provincial database of Lopez-Ruiz & Hasanov (2025, KAPSARC), itself a reconstruction from satellite-observed activity calibrated against national aggregates rather than an official regional account. The eight resource indicators derive from series published at provincial level by GaStat, the Saudi Electricity Company, the Saline Water Conversion Corporation, and the DataSaudi platform. The seven environmental indicators are estimated by apportioning national totals reported by the Ministry of Environment, Water and Agriculture; the procedure is documented in the README and in Section 2.2.1 of the paper. One category of material is not deposited. The national environmental totals and the annual allocation shares from which the seven provincial environmental series were derived are not included: the intermediate files are no longer available in a form that reproduces those series. Every result reported in the paper is reproducible from the provincial series included here, but the derivation preceding them cannot be independently checked. Requirements: Python 3.9 or later with numpy, pandas, and matplotlib.