Understanding Structural Patterns in Private Participation in Infrastructure Projects Using Topological Data Analysis

Private participation bridges the infrastructure investment gap by leveraging private sector expertise and funding. However, infrastructure projects involve multiple stakeholders and carry significant risks, where poor decisions can result in financial losses and system failures. This study examines the World Bank’s Private Participation in Infrastructure (PPI) data using topological data analysis (TDA). While prior research has explored predicting project success, little attention has been given to uncovering underlying patterns in the data. Using techniques such as clustering, manifold learning, and TDA mapper, structural features within the data are examined to identify unique characteristics across different project sectors. Findings show projects grouped by differing investment ranges and investment periods, while outliers stemmed from increased investments or inflation. Variance in the data is explained by time-related economic effects followed by investment effects and inflation effects. Failed projects shared common characteristics, linked to macroeconomic shocks, low GDP, and reduced public sector investments.