Data-driven methods in risk assessment of public-private partnership projects

Purpose Public-private partnerships (PPPs) are an increasingly popular method for delivering public infrastructure projects, but they face significant risks that can impact their success. Effective risk assessment is crucial for mitigating these risks. This study aims to explore the application of data-driven methods, including artificial intelligence (AI) techniques, for risk assessment in PPP projects. Design/methodology/approach This study systematically reviews 30 peer-reviewed journal articles sourced from academic databases, including Scopus and Web of Science. The literature review focuses on the utilisation of traditional statistical models, machine learning (ML) algorithms and initial applications of deep learning (DL) for risk assessment in PPP projects. Findings The findings indicate an increasing trend in employing data-driven approaches for PPP risk assessment. Traditional statistical models, ML algorithms (e.g. random forests and neural networks) and some initial applications of DL have been used. However, the use of advanced DL and natural language processing (NLP) techniques remains limited. Challenges identified include data availability and quality issues, subjective biases, limitations in model specifications and difficulties in generalising findings across contexts. The integration of explainable AI (XAI) methods to enhance the interpretability of complex models is also lacking. Research limitations/implications The findings highlight the need for further research into advanced AI techniques, especially DL combined with NLP and XAI methods, to harness data complexities and deliver transparent risk assessments for PPP projects. Practical implications This research demonstrates the potential of AI techniques to enhance risk assessment in PPP projects. Traditional models often fail to address the complexity of risks, while AI, combined with XAI, provides more transparent and actionable insights. This integration improves decision-making and project outcomes and offers more effective risk management for public and private stakeholders. Originality/value This study provides a novel contribution by systematically reviewing data-driven risk assessment methods in PPP projects, an area with limited exploration of advanced AI techniques like NLP. The research fills a critical gap in the literature by highlighting the need for interpretable AI models to improve the transparency and practicality of AI applications in PPP risk management.