Exploring decarbonization priorities for sustainable shipping: A natural language processing-based experiment

Abstract

The shipping industry is currently the sixth largest contributor to global emissions, responsible for one billion tons of greenhouse gas emissions. Urgent action is needed to achieve carbon neutrality in the shipping industry for sustainability. In this paper, we use natural language processing techniques to analyze policies, announcements, and position papers from national and international organizations related to the decarbonization of shipping. In particular, we perform the analysis using a novel matrix-based corpus and a fine-tuned machine learning model, BERTopic. Our research suggests that the top four priorities for decarbonizing shipping are preventing emissions from methane leaks, promoting non-carbon-based hydrogen, implementing reusable modular containers to reduce packaging waste in container shipping, and protecting Arctic biodiversity while promoting the Arctic shipping route to reduce costs. Our study highlights the validity of NLP techniques in quantitatively extracting critical information related to the decarbonization of the shipping industry.

Keywords

• Shipping
• Decarbonization
• Sustainability
• Natural language processing
• Bertopic
• Machine learning