The Importance of Data Science for Decarbonization Rates in Finance

Global banks develop data platforms or use accredited solution providers, like the carbon disclosure project (CDP), to report their climate impact. Sundeep Reddy Mallu, head of ESG and analytics at Gramener, discusses how more advanced technologies, like natural language processing (NLP) and geospatial imagery, are coming out of the woodwork to integrate greenhouse gas (GHG) emissions into financial reporting and help banks stay on top of sustainability targets. 

While numerous financial institutions have begun to include environmental, social, and governance (ESG)-related matters in annual reports or reporting formats, there are several challenges: 

• Greenwashing
• The lack of standardization of ESG and climate disclosures
• Few private and even listed companies transparently disclose data. 

Regulations will help address these roadblocks to some extent: the US Securities and Exchange Commission (SEC) came out with a climate proposal this year requiring public companies—including banks—to provide audited financial statements with climate-related impacts and GHG emissions. The Corporate Sustainability Reporting Directive (CSRD) in Europe also called for 50,000 companies to follow revised sustainability reporting guidelines, making GHG emissions and ESG data more accessible.

While regulations will democratize and standardize ESG data, data science applications – like artificial intelligence (AI), machine learning (ML), and geospatial analytics – can go further. They can be important enablers for financial institutions to generate faster insights about the impact of investment or lending decisions on sustainability goals and climate commitments.

Financial institutions have only recently woken up to this requirement to reduce decarbonization rates, while sectors that have been under pressure for years regarding environmental performance reporting, like the optimization of assets or logistics operations, have been using data science for a long time. For example, Rio Tinto uses smart mining to optimize the scheduling of mining equipment, while Google uses ML for forecasting wind power generation to make wind farms more profitable. 

With sustainable finance set to play a critical roleOpens a new window in meeting Paris Agreement goals, here’s why data science is essential for our transition to a net zero economy.

Current Challenges with Tracking Decarbonization Rates 

Many financial institutions have committed to net zero through the net-zero banking allianceOpens a new window . And while it looks good on paper, one of the many challenges for financial institutions when tracking decarbonization or refining finance emission numbers is acquiring the data. This depends on the measures they have in place to monitor their customers’ and suppliers’ ecological footprints too. 

While the digitization of information about decarbonization is expected to increase due to government agencies and legislation like SEC and CSRD, much of the available data is currently in an unstructured text format. 

If a financial institution wants financial data for a listed company in the US, it’s easy – it’s all in a digital machine-readable format. If they want greenhouse gas (GHG) data or climate transition plans for any company in their investment portfolio, they have to search for annual reports, sustainability reports, press releases, responses to questionnaires, or other documents. 

This is because financial institutions work with thousands of customers and companies – and, likely, these entities won’t have any emissions data published. Financial institutions can estimate emissions using industry averages, but it becomes difficult to reduce those emissions over time because it’s non-specific. 

Extracting Data from Unstructured Sources with NLP

NLP, a branch of artificial intelligence that deals with understanding written or spoken human language, can be applied in various applications ranging from sentiment analysis to generating automated insights. 

Climate data platforms and ESG rating agencies are increasingly applying NLP to extract climate transition data from unstructured sources, collate data from multiple places, and automate insights from large data volumes. 

NLP can create automated summary reports outlining financial institutions’ climate commitments: It looks at quarterly reports or audio transcripts (unstructured content) and, using ML techniques for automated tagging, sees if they are fulfilling their promises and validates what they’ve expressed publicly.  

Financial institutions can also use NLP to quickly compare portfolio companies’ decarbonization performance, initiatives, and commitments, converting financial metrics to GHG emissions and allowing them to consider climate-related data in investment decisions. For example, by comparing the portfolio company with others in the same sector, NLP could alert investment managers if there’s a climate reporting anomaly—like the exclusion of Scope 3 emissions.

See More: How AI & ML Can Power Advanced Analytics for Corporate Finance

Geospatial Technology for Monitoring Impact in Real-time

Geospatial technology uses computer vision and satellite imagery overlayed with additional layers of data to develop location-specific insights, often for disaster management—but also to help track GHG emissions. 

The continued loss of forests and other natural ecosystems is one of the key risks to achieving the Paris Agreement goal and limiting the temperature rise to 1.5 degrees Celsius. And according to the Global Carbon Budget 2021, changes to land use generate a considerable chunk of emissions yearly.

Banks want to expand their ESG activities and invest in nature-based solutions, but deforestation could unravel their efforts toward decarbonization. Therefore, financial institutions can leverage geospatial technology to monitor deforestation and GHG emissions from portfolio companies and supply chains involved in extracting resources like palm oil, rubber, or soy. 

See More: Managing Data Lifecycle for Financial Services Companies

Banking on Sustainability

High-resolution geospatial imagery is important when verifying the environmental impact of banks’ in-house solutions (as part of their CSR focus) or before banks make funding decisions in projects such as agro-forestry-based plantations or the revival of forest ecosystems. It also means added credibility to ESG metrics and banks’ published financial reporting, which outlines decarbonization rates and supply chain claims. 

Banks are looking to use investment opportunities to drive positive social impact and achieve their sustainability targets, especially post-pandemic. The appropriate technology, like NLP and geospatial imagery, can help banks overcome challenges standing in the way of ESG initiatives through more in-depth reporting and feedback. 

How are you using data science for smarter ESG compliance? Let us know on FacebookOpens a new window , TwitterOpens a new window , and LinkedInOpens a new window .

MORE ON DATA SCIENCE

• How to Pick the Best Data Science Bootcamp to Fast-Track Your Career
• What Is Data Science? Definition, Lifecycle, and Applications
• 5 Ways Data Scientists Can Advance Their Careers