Landslide Prediction Using MATLAB ML and QGIS / PLAXIS

A student-laptop landslide prediction system for a hilly Western-Ghats-type catchment that joins three scales: regional susceptibility mapping with machine learning (logistic regression, random forest, gradient boosting, SVM, MLP) against frequency-ratio and physically based SINMAP baselines, slope-stability analysis of three critical cross-sections with a PLAXIS-style finite-element strength reduction and Bishop's method, and a rainfall early warning that compares an intensity-duration threshold with a machine-learning warning. The random forest reaches a spatial cross-validation AUC of 0.937 and an AUC of 0.918 on an unseen later landslide event; the 5 % 'very high' zone holds 45.1 % of those new landslides. The ghat road cut falls to a factor of safety of 0.56 (FE) / 0.96 (Bishop) at monsoon peak and recovers to 1.17 with drains. A role-based web dashboard shows maps, rainfall scenarios, sections and alerts. Honest scope: the study area and its two landslide inventories are SYNTHETIC (physically generated), and MATLAB, QGIS and PLAXIS are NOT installed on the build PC - their scripts were written but never run; all numbers come from the Python pipeline.

Category: Civil Engineering / Geotechnical Engineering / Landslide Hazard Assessment / GIS and Remote Sensing / Machine Learning / Finite-Element Slope Stability / Early Warning Systems / MATLAB, QGIS and PLAXIS (scripts) with Python

Civil Engineering Project C10 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.