ESG Data Processing

#

In the process of writing my thesis during my training at the IPSSI institution, I had the opportunity to delve into the following topic: The Evolution of ESG Data Processing with the Advancement of AI and Large Language Models.

This topic was selected because it allowed me to leverage the latest advancements in the field of Generative AI and NLP to explore new possibilities for ESG data extraction. Working at Iceberg Data Lab, ESG data is at the heart of the company’s business.

With evolving regulations, publicly traded companies are increasingly required to disclose information regarding their environmental impacts. However, these data are most often made available in unstructured formats (such as PDFs or Excel files), which demands significant effort from analysts to manually extract this data. Hence, the capabilities offered by Large Language Models (LLMs) could potentially revolutionize the way these data are extracted. This is what I attempted to explore through my thesis.

The thesis is available in both English and French (note that the English PDF has not been proofread and may therefore contain some typos, which have likely been corrected in the French PDF).