Predictive Analytics in ESG Investment Evaluation Using Natural Language Processing of Sustainability Reports

Abstract

Environmental, Social, and Governance (ESG) factors have become integral in evaluating corporate sustainability and long-term investment potential. This study investigates the utility of Natural Language Processing (NLP) techniques applied to corporate sustainability reports to predict ESG performance and investment attractiveness. By leveraging sentiment analysis, topic modeling, and machine learning classification models on ESG disclosures, this paper aims to establish a correlation between language features and ESG scores. The predictive analytics framework is validated using ESG data from publicly listed companies between 2018 and 2022. Results indicate that textual indicators in sustainability reports are significantly correlated with third-party ESG ratings, suggesting that NLP can serve as a non-invasive and scalable method for ESG assessment.