<p>This research aims to cluster and characterize Toronto soils based on pressuremeter tests (PMT) by applying machine learning algorithms. More than 400 PMT results were collected from a transit project in Toronto, Canada. Gaussian Mixture clustering was applied to cluster Toronto soils into four main groups with distinct mechanical behaviour: glacial tills I and II, cohesive, and cohesionless deposits. A material index based on the shape of the PMT curve was proposed to enhance the separation between cohesive and cohesionless soils, and PMT soil behaviour charts of material index (M IDvs. Menard modulus (E ) Mere proposed. The isocurves of the coefficient α can be used to estimate soil's elastic moduli by correcting the pre-boring PMT disturbance. Predictive equations for EMandM weID proposed enabling the application of the PMT charts in practice. The application of French PMT charts was investigated, and comparisons between Toronto and UK tills were performed.</p>