Enterprise Scope 1 and Scope 2 Emission Estimation Using Crowd Source Data

Abstract

Globally, GHG emissions have grown by 50% from 1990 to 2018. Even though Enterprises are under significant pressure from investors, consumers, and policymakers to disclose their GHG emissions, only 20% of publicly listed U.S. companies voluntarily disclose emissions data as of now. The major challenge in reporting enterprise emission lies in process/asset level data collection overhead from multiple business units across multiple geographies. To address this challenge, we proposed a Natural language processing (NLP) based framework to estimate Scope1 (S1) and Scope2 (S2) emissions of an Enterprise by learning an emission model from publicly disclosed S1, S2 emissions, and other contextual data such as Environmental, Social and Governance reports. This AI based tool will help small and medium enterprises to estimate their S1 and S2 emissions without any primary data.