using machine learning techniques to analyze ESG reports: what machine learning can uncover in ESG reports to auton the European sustainability reporting standards

Rodel, Kevin

http://hdl.handle.net/10362/201937

Utilize este identificador para referenciar este registo.

Nome:	Descrição:	Tamanho:	Formato:
2024_25_Fall_59407_Kevin_Roedel_REFINED.pdf		5.25 MB	Adobe PDF	Ver/Abrir

Contacte-nos

Autores

Rodel, Kevin

Orientador(es)

Silva, Nuno André da

Wagenknech, Tobias

Resumo(s)

This work explores the application of natural language processing (NLP) to analyze Environmental, Social, and Governance (ESG) reports, addressing challenges in analyzing and benchmarking ESG efforts. The developed ESG Analyzer extracts ESG topics and trends on a sectoral, industrial, and country level and the ESRS Analyzer provides insights into the reporting of European Sustainability Reporting Standards (ESRS). Despite data constraints, the models achieved accuracies of 88% and 82%. The findings highlight variability in ESG practices, offering actionable benchmarks and insights, accelerating report analysis. By advancing these analytics, this study contributes to the understanding of ESG trends and the adaptation of ESRS regulation.

Palavras-chave

Environmental Social Governance (ESG) Sustainability Reporting Disclosure Benchmarking Corporate sustainability Reporting directive (CSRD) European Sustainability Standard reporting (ESRS) Machine learning (ML) BERT ESG BERT model Natural language processing (NLP)