Analysing community-level spending behaviour contributing to high carbon emissions using stochastic block models

Area of Science:

• Data Science
• Environmental Science
• Network Analysis

Background:

• Financial transaction datasets are increasingly utilized for estimating carbon emissions from consumer spending.
• Understanding consumer spending behavior is crucial for effective carbon reduction strategies by policymakers and financial institutions.

Purpose of the Study:

• To develop a method for identifying consumer spending patterns within large financial transaction datasets.
• To enable targeted carbon reduction strategies by linking spending patterns to high-emission merchant categories.

Main Methods:

• Utilized stochastic block modelling for community detection on a bipartite network.
• Developed a modified weighting system for the bipartite network to maintain constant average community spending across categories.

Main Results:

• Successfully formed communities of consumers with similar spending patterns.
• Identified specific merchant categories contributing to high carbon emissions for distinct consumer groups.
• Demonstrated a framework for analyzing the interplay of spending, energy usage, and transportation habits.

Conclusions:

• Transaction datasets and stochastic block modelling offer valuable insights for financial institutions aiming to decarbonize.
• The study provides a controlled framework for researchers to analyze the impact of consumer behavior on carbon reduction efforts.