Data and code for feasibility-bounded diet optimization across ten countries

Analysis code and derived datasets for a feasibility-bounded diet optimization study across ten countries (4.15 billion people). The repository contains Python scripts, processed result tables, and reproducibility documentation for three phases of analysis: baseline dietary assessment, linear-programming optimization under plus-or-minus 50% intake-change bounds, and sensitivity and uncertainty analyses. Reported outcomes include DALYs averted and environmental indicators, including greenhouse gas emissions, land use, freshwater withdrawals, and eutrophication. The repository also includes region-specific FAO GLEAM v3.0 ruminant emission coefficients used in the environmental analyses. Upstream source datasets, including GBD 2023, Global Dietary Database 2018, and FAO Food Balance Sheets, are not redistributed because they are provided by third-party sources and may be subject to licensing or access restrictions; acquisition details are described in the accompanying README.