Assessing the interrelationships among SDG 6-related indicators using explainable machine learning

Abstract

 The United Nations indicated the Sustainable Development Goals (SDGs) as goals to be achieved by 2030. However, due to the complexity of SDGs, there are various interrelationships (synergies and trade-offs) among the goals and targets, which need to be accurately understood in order to improve the progress of SDGs. In this study, we focused on Goal 6 "Clean Water and Sanitation" and investigated the interrelationships of Goal 6 and other goals using explainable machine learning. Totally, 80 indicators were selected for 176 countries, and we used cluster analysis and LIME (Local Interpretable Model-agnostic Explainations) to examine the interrelationships between Target 6.1/6.2-related indicators (population with access to drinking water and sanitation) and other SDGs indicators. The results show that there are synergistic relationships between Target 6.1/6.2-related indicators and other indicators such as electricity, poverty rate, and mortality, and a trade-off with energy consumption and waste. This study proposes the use of explainable machine learning as a method for quantitatively analyzing the interrelationships among various Goals and Targets.