Abstract: Data-driven models using historical project attributes to estimate future construction costs, such as multiple-regression analysis and artificial neural networks are both proven techniques that highway agencies could adopt ...