Purpose This study aims to investigate the influence of climate change and socio-economic parameters on the scale of infrastructure projects using Artificial Intelligence (AI)-based Machine Learning (ML) algorithms. By focusing on drinking water systems in Adana, Türkiye, the research provides a robust framework for forecasting infrastructure demands under varying environmental and demographic scenarios. Design/methodology/approach A two-stage prediction framework was developed. Stepwise assessment ratio analysis and principal component analysis (PCA) methods were employed for feature selection from 15 independent variables. Supervised ML algorithms – Ridge, support vector regression (SVR), Least Absolute Shrinkage and Selection Operator and XGBoost – were used to estimate distribution flow rates and infrastructure project lengths. Models were validated using hold-out and k-fold cross-validation techniques. GridSearch was applied for hyperparameter tuning, and the Aydeniz climate classification was integrated to reflect drought impacts. Findings The SVR and ridge algorithms showed the highest predictive accuracy, with Nash–Sutcliffe efficiency and R2 values exceeding 0.85. Socio-economic factors, particularly population and number of subscribers, were found to be more influential than climatic variables. The models successfully predicted future infrastructure project sizes with low error margins, demonstrating strong generalizability and practical reliability. Research limitations/implications The study is based on a regional case (Adana, Türkiye), which may limit the generalizability of the model to other geographic areas without retraining. Additionally, access to long-term, high-resolution climate projections may further enhance model robustness. Practical implications The proposed AI-based models support decision-makers in prioritizing infrastructure investments, managing water resources and adapting to climate risks. The findings can guide municipalities and planners in resource allocation and long-term urban infrastructure development. Social implications Ensuring reliable water infrastructure in the face of socio-environmental stressors contributes directly to public health, economic resilience, and climate adaptation. The model promotes sustainable urban planning aligned with global development goals. Originality/value This study uniquely integrates Aydeniz climate classification with advanced ML to predict infrastructure project size. It introduces a data-driven, scalable approach for estimating long-term needs in water infrastructure under socio-environmental uncertainty. The methodology advances infrastructure planning through scientific and computational innovation.
Baltacı et al. (2026) studied this question.