Promoting a New Sustainable Metric to Measure Progress: A Data Science Approach

This thesis challenges the limitations of using Gross Domestic Product (GDP) as the primary indicator of national progress, emphasizing that GDP focuses solely on economic output and overlooks the complexity of sustainable development. Specifically, this study follows the path traced by the Beyond GDP Initiative, and consequently, its central goal is to create a composite index that integrates economic, social, and environmental dimensions. Moreover, the different stages of development of countries are innovatively the focus point of the research. Indeed, the scope of the metric is also to adapt the weighting of its components based on a country’s level of development, ensuring a more accurate and comparable reflection of national progress. Additionally, this is done in order to allow the new composite metric to promote tailored action plans in order to maximize the economic and ecological country’s performance for each stage of development of a country. A major contribution of this thesis is the rigorous mathematical analysis of potential bias through the application of machine learning and neural network models. Specifically, machine learning algorithms are fitted to each component of the newly proposed composite metric using independent datasets. This is to ensure that each component (economic, social, ecological) does not exhibit bias toward any different socio-economic or ecological category and guarantees that their linear combination can allow comparative equity. This approach mirrors what in literature is done in many fields, including finance and healthcare, where machine learning models are fitted on individual components to ensure bias-free results before combining them into a composite metric. In the sustainable development context, a comparable evaluation means assessing countries’ performance without favoritism, ensuring that comparisons are based solely on objective and universally accepted criteria. This type of analysis, commonly done in the social domains, is for the first time applied in the sustainable development field, marking another novel aspect of the research. The bias testing analysis lays the groundwork for applying the metric to rank nations according to a new composite index of sustainable development, offering insights on sustainable trajectories from 1990 to 2020. This research represents the foundation of a broader project. Indeed, the metric evaluation could be expanded by conducting a detailed analysis of how the weighting of its components could change the sustainable progress outcomes.