Proj S11 – ESGpedia – Sustainability Intelligence Platform

An automated pipeline that scrapes, extracts, and benchmarks ESG metrics from corporate sustainability disclosures across Southeast Asia — targeting 85%+ accuracy across 1,100+ companies.

Introducing Proj S11 – ESGpedia – Sustainability Intelligence Platform

Our project delivers a scalable sustainability intelligence platform that automatically extracts and benchmarks ESG metrics from corporate sustainability disclosures across Southeast Asia, covering over 1,100 companies across Singapore, Malaysia, and Indonesia. Powered by the QWEN Vision Language Model and advanced anomaly detection, the platform achieves up to 86.1% extraction accuracy — empowering investors and analysts to benchmark ESG performance across industries without the need for manual report review.

Team members

Ng Yu Hueng (ISTD), Lam Yu En (ISTD), Khoo Teng Jin (ESD), Zachary Low Yang Kai (ISTD), Chia Chun Mun (ISTD), Oon Shao Ren (ISTD)

Instructors:

Zhao Na

Writing Instructors:

Belinda Seet
Dominic Edmund Kim San Quah

The Problem Statement

How might we design a scalable sustainability intelligence platform that automatically scrapes, extracts, and benchmarks ESG metrics from diverse corporate disclosures across Southeast Asia, achieving over 85% extraction accuracy while enabling consistent benchmarking across 26+ indicators and over 1,100+ companies?

Our Solution

Platform Features

Data Validation

Extracted values for 305 companies were compared against ESGpedia’s curated dataset, then individually verified against original PDF reports to understand root causes of discrepancies.

Data Quality

ESG data is inherently noisy — values span orders of magnitude, units vary widely, and many companies disclose only partial data. A four-layer detection system catches errors that any single method would miss.

Acknowledgements

We extend our sincere gratitude to our industry mentors, Mr. Benjamin Tan and Mr. Jin Ser, for granting us the opportunity to collaborate on this project with ESGpedia. Their unwavering support, openness to our ideas, and insightful guidance have been pivotal in shaping the project’s direction.

We are deeply grateful to our SUTD mentor, Prof. Zhao Na, whose consistent guidance and technical expertise kept us on track throughout the project.

Our heartfelt thanks go to Mr. Dominic Quah for his invaluable advice on improving our presentations and reports, as well as guidance on presenting ourselves effectively to both our industry partner and SUTD.

Finally, we acknowledge SUTD for their administrative support, which played an essential role in facilitating our project’s development.