Estimating carbon footprints from large scale financial transaction data

• Published in: Journal of industrial ecology Vol. 27; no. 1; pp. 56 - 70
• Main Authors: Trendl, Anna, Owen, Anne, Vomfell, Lara, Kilian, Lena, Gathergood, John, Stewart, Neil, Leake, David
• Format: Journal Article
• Language: English
• Published: New Haven Wiley Subscription Services, Inc 01.02.2023
• Subjects: big data; Carbon; Carbon footprint; Consumption; Customers; Data capture; Data quality; Data sources; Emission measurements; Emissions; Environmental impact; Footprint analysis; greenhouse gas emissions; Household expenditure; industrial ecology; Polls & surveys; Profiles; Qualitative analysis; Standard data; Surveys; Transactions; United Kingdom > UK
• ISSN: 1088-1980; 1530-9290
• DOI: 10.1111/jiec.13351

Financial transactions are increasingly used by consumer apps and financial service providers to estimate consumption‐based carbon emissions. This approach promises a low‐resource, ultra‐fast, and highly scalable approach to measuring emissions at different levels of potential policy intervention—spanning the national, subnational, local, and individual level. Despite this potential, there is a lack of research exploring the validity of this approach to carbon profiling. Here we address this oversight in three ways. First, we provide a step‐by‐step description of our approach toward estimating carbon footprints from micro‐level transaction data generated by more than 100,000 customers of a large retail bank in the United Kingdom. Second, we quantitatively compare emission estimates obtained from transaction data with those calculated from a more standard data source used in carbon profiling, the largest household expenditure survey in the United Kingdom. Third, we offer a detailed qualitative comparison of the advantages and disadvantages of transactions versus alternative data sources (such as survey data), across key dimensions including data availability, data quality, and data detail. We find that financial transactions offer a credible alternative to survey‐based sources and, if made more widely accessible, could provide important advantages for profiling emissions. These include objective, micro‐level data on consumption behaviors, larger sample sizes, and longitudinal, frequent data capture.