VOGOSEN

Aligning ESG data using NLP

ESG data is getting more complex with diverse structured and unstructured sources. These sources relate to multiple indicators, ESG dimensions, and other standard classifications such as SDGs and SASB material aspects
Published by Zakaryae Boudi on 22 Jul 2021 · Keywords: aiesgnlpsasbsdgsvogosen
ESG data is getting more complex with diverse structured and unstructured sources. These sources relate to multiple indicators, ESG dimensions, and other standard classifications such as SDGs and SASB material aspects. 

Why is it important to align ESG data?

Because investors need to build the most factual understanding of ESG aspects across their investment universe: at the same time, a single perfect data source for every investment case does not exist. In reality, many providers offer different information from complementary angles. Capturing the full ESG picture of each specific investment case requires a holistic approach to processing all the data available. That is also why some investors develop in-house ESG scoring / analysis methodologies built on various data sources.
Natural Language Processing (NLP) is a field of Artificial Intelligence (AI) that can help align ESG data elements in an efficient way. Let us imagine you want to develop a custom ESG scoring approach based on SASB material aspects, which takes into account the scores provided by two different agencies and presented as a taxonomy of indicators. Deploying a small NLP program not only helps getting the execution faster, but also makes the whole project scalable to more data sources.
In particular, the NLP system will let you know which indicators in your data are actually material to a specific company. For example, if "Safety of Clinical Trial Participants" is a material issue to a company in your portfolio, and your data shows scores across many different indicators, the AI will instantly highlight the semantically relevant indicators you should focus on, such as "Clinical Trial Standards" or "Health & Safety Certifications". Based on that, it is possible to automatically interface data elements from different sources, before computing internal ESG scores with proprietary formulas.
An example of mapping indicators with SASB material aspects
An example of mapping indicators with SASB material aspects

What can you do now?

You would like to know more about AI for ESG and look forward to exploiting NLP for your ESG data? Let us know and get in touch. We will be happy to discuss!
We help our clients use advanced natural language analytics to turn ESG data into practical investment insights.
Photo of Zakaryae Boudi

Zakaryae Boudi

As an entrepreneur and consultant, Zakaryae Boudi developed an in-depth expertise in multiple areas including new technologies (AI, Big Data, Cloud computing, etc.), public policies and international development (employment / training, security, transport, infrastructure, sustainable development, PPP, etc.).
Before launching one of the leading labor market intelligence start-ups in France and French-speaking Africa, Zakaryae had the opportunity to work in international collaborative teams on diverse economic and social challenges. In particular, he led the development of the global PPP and financing centre of excellence at the International Union of Railways, as part of the Sustainability Unit, and managed a major digital program focused on talents and skills.
Zakaryae holds an engineering degree from the Ecole Mohammadia d'Ingénieurs (Morocco) as well as a Master from Sciences Po Paris, and he recently completed a PhD thesis on AI systems' control using mathematical formal methods.