Translating Data into Insights: New Technology Enhances ESG Data Analysis and Investing

Overview

During a visiting scholar appointment at Stanford University’s Precourt Institute for Energy (March 2019 – September 2020), this project explored how machine learning and natural language processing could transform the way investors and researchers understand Environmental, Social, and Governance (ESG) factors.

The core question: can we use technology to cut through the noise of ESG reporting and find the signals that actually matter for sustainable investing?

Research Approach

We conducted a large-scale bibliometric analysis of 26,111 academic articles from Web of Science (1973–2019) to map the evolution of ESG research. Using NLP techniques, we identified emerging themes, tracked how the field evolved from early corporate social responsibility studies to today’s data-driven sustainable finance, and highlighted gaps where technology could add the most value.

The research also explored practical applications of ML for ESG data analysis, including automated classification of ESG-relevant information and assessment tools for corporate sustainability performance.

Key Outcomes

The work resulted in a publication in Corporate Social Responsibility and Environmental Management (2023), co-authored with In S.Y. and Eccles R.G. The paper provided a comprehensive technology-driven perspective on the ESG research landscape and identified opportunities for AI-enhanced sustainable investing.

Significance

This project bridged the gap between academic ESG research and practical investment applications, demonstrating how NLP and machine learning could make ESG analysis more systematic and objective — a topic of growing importance as sustainable finance regulations expand worldwide.

Collaborators

• Stanford Precourt Institute for Energy: Host institution
• Research Team: In S.Y., Lee Y.J., and Eccles R.G.