Code and demonstration data for « Geostatistics versus machine learning for reconstructing aquifer basement geometry from a dense legacy VES archive (Abadla plain, SW Algeria)

Reproducible Python code and a synthetic demonstration dataset accompanying the manuscript. The code implements ordinary kriging, random forest, gradient boosting, regression kriging, and random vs spatial cross-validation, together with the survey-thinning experiment and figure scripts. The real vertical electrical sounding data are the property of the ANRH (Algeria) and are not redistributed; a synthetic dataset reproduces the full workflow. Reproducible Python code and a synthetic demonstration dataset accompanying the manuscript. The code implements ordinary kriging, random forest, gradient boosting, regression kriging, and random vs spatial cross-validation, together with the survey-thinning experiment and figure scripts. The real vertical electrical sounding data are the property of the ANRH (Algeria) and are not redistributed; a synthetic dataset reproduces the full workflow.