Application of Statistical and Machine Learning Methods for Environmental Monitoring Based on ESG Data

Abstract

This paper examines the application of machine learning methods to environmental monitoring based on ESG data using the case of Ukraine, with a comparative analysis of the European Union. The study analyzes the relationships between energy use, renewable energy share, and $\text{CO}_{2}$ (greenhouse gas) emissions. The analysis is based on World Bank data for Ukraine and aggregated Eurostat data for the EU. Correlation analysis revealed a strong positive relationship between energy use and emissions, and a negative relationship between renewable energy share and environmental pressure. Linear regression models showed high explanatory power for both datasets, with $R^{2}$ values above 0.97. The Random Forest algorithm was additionally applied to assess factor importance and showed that renewable energy share is the most influential variable in both cases. The results suggest that Ukraine's environmental progress will largely depend on expanding renewable energy and aligning with EU environmental policy.