Data & Code for Sparky BA and AF Model Training and Evaluation

This repository contains the scripts and supporting data (csvs and netcdfs) used to compile, train, run and evaluate the Sparky-BA and Sparky-AF modelling systems described in the publication “Socioeconomic Development Masking the Climate Signal in Global Fire Trends”. All datasets included here are fully referenced within the manuscript, and users should credit the original data providers where appropriate. The repository also includes the plotting routines used to produce the figures presented in the paper. Key processed output data required to reproduce the study’s results are included. However, due to a combination of data-sharing restrictions and file-size limitations, the complete raw dataset cannot be hosted here. Authors will provide access to any additional data that can be shared but is not included in this repository, please contact us directly for details.