Data architecture for: Biodiversity loss of the Brazilian soy market

This repository contains the direct and primary deposit of the data directory architecture required for the study analysing biodiversity loss in the Brazilian soy supply chain. This record is a standalone data repository hosting the heavy computational assets, specifically the data folders that cannot be stored in GitHub due to file size constraints. To ensure open science and computational transparency, this repository acts as an optimised validation dataset. It contains the complete data structures, tabular databases, spatial boundaries, and intermediate outputs required to execute, test, and verify the model's analytical pipeline for a default reference year (2019) and a single reference municipality. Deposited Data Files This repository hosts two core compressed archives that contain the original directory structures required by the R modeling environment: Input_data.zip: The primary data folder containing all raw input assets, including spatial biophysical parameters, vector shapefiles, and localized tabular databases necessary to initialize the simulation. Output_data.zip: The baseline output folder containing intermediate computational layouts and the .gitkeep anchor file required by the workflow to log and save optimization results. How to Use This Dataset The files hosted in this repository are designed to populate the data directories of the companion computational model: Download the Data: Download both Input_data.zip and Output_data.zip directly from this Zenodo record. Extract: Unzip both files. This will generate two distinct folders named Input_data and Output_data. Integration: Place both extracted folders directly inside the root directory of your local RStudio project environment (parallel to the master .Rproj file) where the execution scripts are located. Execution: Once the data folders are in place, the relative paths within the R scripts will automatically detect the assets and initialise the master simulation pipeline. Protocols for Full-Scale Historical Data Access The complete, multi-terabyte historical time-series datasets (covering 2004-2022) and high-resolution continental raster layers omitted from this validation archive due to platform storage boundaries or proprietary licensing restrictions are fully cataloged within the manuscript's methodological appendix. For access protocols to the comprehensive macro-scale core data layers, please contact the corresponding author as indicated in the primary publication.