Date of this Version



environmental sustainability, unsupervised learning, self-organized maps, global analysis, environmental performance index, clustering framework


The importance of sustainable development has risen in recent years due to the significant number of people affected by lack of access to essential resources as well as the need to prepare for and adapt to intensifying climate change and rapid urbanization. Modeling frameworks capable of effectively assessing and tracking sustainability lie at the heart of creating effective policies to address these issues. Conventional frameworks, such as the Environmental Performance Index (EPI), that support such policies often involve ranking countries based on a weighted sum of a number of relevant environmental metrics. However, the selection and weighing processes are often biased. Moreover, the ranking process fails to provide policymakers with possible avenues to improve their country’s environmental sustainability. This study aimed to address these gaps by proposing a novel data-driven framework to assess the environmental sustainability of countries objectively by leveraging unsupervised learning theory. Specifically, this framework harnesses a clustering technique known as Self-Organized Maps to group countries based on their characteristic environmental performance metrics and track progression in terms of shifts within clusters over time. The results support the hypothesis that the inconsistencies in the EPI calculation can lead to misrepresentations of the relative sustainability of countries over time. The proposed framework, which does not rely on ranking or data transformations, enables countries to make more informed decisions by identifying effective and specific pathways towards improving their environmental sustainability.


DOI: 10.3390/su12020563