Data-driven Risk Modeling for Infrastructure Project Delivery to Mitigate Schedule and Cost Overruns

Abstract

Despite the development of digital systems and project controls, infrastructure projects are still prone to uncontrolled risks, delays in schedules, and higher costs. This scoping review explores how data-driven risk modelling can be used to make predictions, quantify, and manage schedule and cost risks in infrastructure delivery environments. Scopus, Web of Science Core Collection, and Engineering Village were used to find studies published between 2015 and 2025 using a PCC-framed question and PRISMA-ScR-managed process. A standardised template was then used to screen and chart them. A total of 19 empirical studies were synthesised, all of which were highways, roads, railways, metro systems, marine works, irrigation, educational infrastructure, utilities, and portfolios of public infrastructure, in developed and developing contexts. The approaches reviewed involved machine learning, Bayesian networks, Monte Carlo simulation, with hidden Markov models, earned value-based forecasting, fuzzy systems, and hybrid analytics, and were tested against cost overrun, delay, contingency, and risk outcomes. Although data-driven and probabilistic models improve decision support and prediction across most applications, wider adoption is constrained by issues such as limited validation beyond single settings, weak integration with live control systems and poor data quality. This review provides coherent evidence, which has been fragmented around identified priorities of governance, validation, and integration.