Solar energy potential assessment and investment profitability analysis: The case of North Macedonia

This study presents a novel integrated framework that combines meteorological data analysis, machine learning forecasting models, clustering techniques, and economic modeling to offer the first data-driven investment roadmap for solar energy in North Macedonia. High-accuracy solar radiation forecasts for 2025 were generated using the Gradient Boosting Regression model (R² = 0.953) using NASA POWER data for the period 2020–2024. Cities were grouped into clusters based on their climatological characteristics; cities such as Štip, Veles, and Kavadarci demonstrated the highest solar potential and economic efficiency. Economic analyses identified payback periods of 3–5 years and average ROI (Return on Investment) values ​​exceeding 24%. It was determined that increasing electricity prices linearly affected the ROI, reaching 33% with an increase in price to $0.12/kWh. The findings highlight the importance of a new integrated decision-support framework merging ML-based solar forecasting, spatial clustering, and dynamic economic modeling tailored for data-limited emerging markets such as North Macedonia