Towards Global Sustainability: Exploratory Analysis Through Unsupervised Machine Learning Techniques

In the context of sustainable development worldwide, issues related to public policy for building a sustainable world have played a significant role in advancing sustainable, economic, and technological aspects with high potential. In this scenario, this study is based on a consolidated database from two different sources of information, where its integration comprises 19 socioeconomic and sustainability variables across 145 countries. Using unsupervised machine learning techniques, an exploratory analysis was conducted based on statistical modelling for constructing clusters among countries with similar characteristics in global sustainable development. The study contributes to elucidating the correlations between variables, identifying those that exert more significant influence and impact regarding sustainable assessment scenarios. Implementing the k-means algorithm, associated with computational support, resulted in constructing eight clusters as an ideal configuration to represent the sustainability context in the world. This made it possible to identify countries with similar quantitative characteristics concerning sustainability. Finally, the conclusions from the results and the limitations identified in this study are presented. In addition, proposals for future research are outlined, aiming to explore sustainability issues in depth globally and overcome the gaps identified in this analysis.