Information Extraction: Unstructured to Structured for ESG Reports
Zounachuan Sun, Ranjan Satapathy, Daixue Guo, Bo Li, Xinyuan Liu, Yihan Zhang, Cheng-Ann Tan, Ricardo Shirota Filho, Rick Siow Mong Goh · 2024
The diverse ESG reporting standards adopted worldwide lead to a significant increase in the volume of unstructured reported information, bringing the need for more efficient processing and standardization of ESG information. Our study reveals a significant enhancement in the comprehensiveness of ESG information disclosed by the real estate industry in Singapore, following the SGX’s ESG reporting guidelines. We improved efficiency in collecting and standardizing ESG metrics from reports using an NLP-based automatic extraction algorithm. This project not only streamlines the ESG data extraction process but also contributes to the broader goal of converting unstructured data into structured formats. Furthermore, it sets a valuable precedent for the industry, fostering increased transparency and accountability within Singapore’s corporate landscape and potentially influencing global standards. Using the automatic extraction technique, we are paving the way for a more informed and responsible approach to corporate sustainability reporting.