On 1-2 Dec 2021, OS-Climate (OS-C) held technical deep dive sessions about the OS-C platform and tools. In the session “Data Extraction: Climate Metrics & Natural Language Processing”, Ismail Demir, Data Scientist, IDS – together with Lea Deleris, Head of RISK Artificial Intelligence Research, BNP Paribas, Jeremy Goh, Data Scientist, BNP Paribas, and Karan Chauhan, Data Scientist, Red Hat – presented the Natural Language Processing (NLP) toolkit.

IDS uses the NLP Toolkit to extract KPIs (Key Performance Indicators) from PDF documents. The PDFs are sustainability or annual reports published by any company, and the KPIs are ESG-relevant metrics, e.g. Direct Greenhouse Gas Emissions.

The toolkit consists of two complimentary components:

  • A rule-based component to extract KPIs from tables
  • A machine learning component to extract KPIs from body text

IDS experts are proud to collaborate in the development of the algorithms allowing for a faster and easier ESG data sourcing. This will give asset managers and asset owners a greater ability to measure and manage their physical and transition risks.

Watch here