VOGOSEN

Data Mining and Apps in AI-Powered ESG Information System

We conclude our AI in ESG concept series with a couple of more concrete subjects: data mining architecture and application design.
Published by Jiulin Teng on 12 Dec 2021 · Keywords: aiappbig-datadata-miningesgvogosen
Having discussed the necessity, capability potential, and design and deployment principles in previous posts, we conclude our AI in ESG concept series with a couple of more concrete subjects: data mining architecture and application design.

Data Mining

Efficiently aggregating a large number of data points starts with the implementation of a scalable web-mining system able to capture—in real time—all ESG related information publicly available on the Internet.
Such a system must operate within a Big Data architecture that is adapted to the size and dynamics of data as well as the targeted web sources. Its architect needs to take into account security and data integrity, in particular when combining cloud infrastructure and local data centers. This process involves guaranteeing real-time mining when needed, ensuring the optimization of the database structure, adapting web-crawlers to site structures (targeted and exploratory crawling), incorporating metadata, and pre-processing functions. The outputs can then feed the analytical systems in place (specialized analytics for ESG research, ESG internal apps, etc.). We can also use these feed data to (re)train the AI modules.

App Design

The AI-powered ESG information system can be deployed as pure APIs upon which investor’s in-house development team build their proprietary interfaces. Alternatively, this information system can come with application interfaces built to the specifications of the investor on an Agile basis.
In either case, the user interface should be designed and developed progressively, in the same spirit of the progressive deployment of the entire ESG Information System. That is, while in the beginning a more structured user interface is preferred, as the system evolves it will be possible to enrich the user interface with AI, too. For example, investment managers will be able to ask questions to the information system directly, and the AI will identify the appropriate module to seek the corresponding answer.
The ESG Information System can also be integrated with software and information systems that other parts of the organization use.

To read more on AI in ESG, download our White Paper here.
Photo of Jiulin Teng

Jiulin Teng

Jiulin Teng has deep understanding of key topics in management science and new technologies. He also has broad knowledge in macroeconomics, business policy, and software engineering.
Upon completing his dissertation on the application of game theory in competitive strategy, Jiulin Teng worked as a consultant at McKinsey & Company (Belgium). More recently, he has launched and participated in a number of startups in blockchain, IT, and AI. He is a proficient database architect and is a full-stack developer with good command of modern programming languages. He has also taken an interest in Natural Language Processing.
Jiulin Teng holds a PhD in Strategic Management and Business Policy from HEC Paris (France). He also has a Master’s degree in Bioentrepreneurship from Karolinska Institute (Sweden) and a Bachelor’s degree in Biological Sciences from Nankai University (China).