Machine learning for spatial disaggregation of regional transport data in the EU

The European Union (EU) is actively working to combat climate change and promote sustainable development by reducing greenhouse gas (GHG) emissions. The transport sector, a major contributor of GHG emissions, was at the forefront of these initiatives. After experiencing steady growth from 2013 until 2019, there was an abrupt decrease in 2020 due to the COVID-19 pandemic. However, preliminary estimates indicated a rebound of 7.7% for transport emissions in 2021, according to the [Agency (2021)]. Nonetheless, further research is necessary in order to devise effective strategies for regional decarbonization within this challenging sector. An analysis of the transport sector in Europe reveals significant disparities in emission trends across different regions. According to [Eurostat (2021)], Western European countries have generally experienced greater decreases in transport emissions compared to Central and Eastern European nations, which have made slower progress. Furthermore, the European Environment Agency [Agency (2021)] points out that urban areas tend to have higher emissions due to higher population densities and greater demand for transportation. These discrepancies underscore the necessity for spatial disaggregation when developing tailored decarbonization strategies for different regions. To address the intricacies of regional decarbonization potentials, this research aims to apply machine learning techniques to enhance the accuracy of estimating transport-related metrics at a regional level. This, in turn, will facilitate the identification of decarbonization opportunities within the transport sector. More precisely, the study seeks to establish a framework that utilizes machine learning methodologies for spatial disaggregation, a critical process for understanding the factors that influence emissions on a regional scale and devising efficient mitigation strategies.