Commercial supersonic passenger transport has been absent from global aviation for more than two decades, largely due to regulatory, geographic, and economic constraints. While renewed interest in supersonic travel has emerged with advances in aircraft design, there remains a lack of scalable methods for assessing where such operations could be viable. This study evaluates supersonic feasibility at the route level using a data-driven framework that integrates engineering, regulation, and economics. A global dataset comprising 435 city-pair routes was constructed using aircraft performance estimates, great-circle routing, over-water routing fractions, and demand indicators derived from population and gross domestic product data. Routes were labeled as feasible or unfeasible based on domain-informed criteria, and supervised machine-learning models were trained to learn a continuous feasibility score between 0 and 1. A Decision Tree classifier was used to extract interpretable feasibility rules, while an Extreme Gradient Boosting (XGBoost) classifier provided predictive performance. Model behavior was analyzed using SHapley Additive exPlanations (SHAP). The results show that over-water routing fraction is the dominant determinant of feasibility, followed by time savings and great-circle distance, with demand contributing in marginal cases. The framework produces a ranked set of candidate routes as well as a predictive engine for future routes.