A Multi-Decadal Curated Geospatial and Hydro-Climatic Benchmark Dataset for Desert Locust Outbreak Analysis in Egypt

Overview This repository contains the official, machine-learning-ready benchmark dataset and technical validation codebase for the manuscript: "A Multi-Decadal Curated Geospatial and Hydro-Climatic Benchmark Dataset for Desert Locust Outbreak Analysis in Egypt"The project provides a rigorous framework for stress-testing predictive modelling pipelines against ecological risk and climate-driven covariate shifts over a 44-year timeframe (1981–2024). Repository Contents 1. Desert Locust Outbreak Analysis in Egypt Dataset.csv: A curated 21-dimensional geospatial and hydro-climatic feature matrix. It fuses historical field records from the FAO Locust Hub with high-resolution reanalysis data from ECMWF ERA5-Land, NASA POWER, and NASA MODIS (MOD13A3) NDVI archives.2. paper_data.py: The full Python execution script containing the technical validation pipelines. It includes Exploratory Data Analysis (EDA) visualizations, feature importance ranking, learning curve evaluations, temporal split sensitivity benchmarks, and spatial-block statistical significance bootstrap testing using Scikit-Learn and XGBoost frameworks.3. requirements.txt: A text file pinning the exact Python package versions required to reproduce the study and avoid execution discrepancies.4. README.txt: Complete guidance on installation prerequisites and replication steps. Use and Licensing The curated data and validation code are released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. The underlying raw source repositories (FAO, ECMWF, and NASA) carry their own separate usage terms which apply independently.