Predictive Analytics for Improving Project Delivery Efficiency in Construction and Infrastructure Management
Mercy Amuna *
Project Management, Northeastern University, ME, USA.
Kelvin Ebo Rabbles
College of Professional Studies, Northeastern University, USA.
Jude Nartey Beantey
College of Professional Studies, Northeastern University, USA.
Moyosore Ikmat Oduola
School of Computing, Engineering and The Built Environment, Edinburgh Napier University, Scotland, United Kingdom.
Stephen Okyere Boansi
Roux Institute, Northeastern University, Maine, USA.
*Author to whom correspondence should be addressed.
Abstract
Construction and infrastructure projects continue to experience unmanaged risks, schedule delays and persistent cost overruns, despite advances in digital tools. This scoping review examines the application of predictive analytics to forecast project delays, performance outcomes and risks in real infrastructure and construction settings. Using a PCC-framed question and a PRISMA-ScR-guided process, studies published between 2015 and 2025 were identified in the ASCE Library and Scopus. These were then charted and screened in duplicate using a standardized template. Nineteen empirical studies were synthesized, covering urban roads, marine works, tunnels, highways, prefabricated construction and buildings, in both developed and developing contexts. These models include artificial neural networks, support vector machines, metaheuristic-optimized hybrids, tree-based ensembles, regression baselines, and gradient boosting, which have been validated against risk outcomes, project-level cost and schedule. Across most use cases, hybrid and machine learning models outperform conventional regression; however, deployment is constrained by integration with existing project controls, organizational capabilities and data quality. This review proposes a thematic structure that links model families to decision use cases, consolidates dispersed evidence, and highlights priorities for data governance, validation, and system integration. These insights offer a practical roadmap for project organizations and researchers seeking to embed predictive analytics in infrastructure management and mainstream construction.
Keywords: Predictive analytics, construction project management, machine learning, cost and schedule performance, risk forecasting