Global Seismic Landslide Prediction Application using XGBoost

Seismic landslides have historically been a major natural hazard, causing high fatality rates and significant economic damage. As a result there is a need for the spatial prediction of seismic landslides, facilitated by statistical and machine learning (ML)–based models. This dataset represents a global prediction model for seismic landslides that is applicable worldwide. The model is developed using the eXtreme Gradient Boosting (XGBoost) algorithm and allows near-real-time landslide predictions. The dataset includes a user-friendly application script that generates and visualizes prediction of landslides using input data from USGS EarthExplorer and NASA EarthData, along with soil thickness and soil type data from SoilGrids.