Algorithm Update: How our Carbonizer just got better
January 4, 2021
We're updating our Carbonizer to improve the accuracy of your carbon footprint estimate.
Joro is launching an improved Carbonizer, the algorithm we use to convert a user’s credit card transactions into a personal carbon footprint. This update will improve carbon footprint estimates for users via two main changes:
1. Integrating a new, more granular and rigorous dataset. Previously, Joro constructed our transaction categories and carbon weights using a combination of top-down and bottom-up estimates from product-based Life Cycle Assessments (LCAs). Our updated methodology uses a rigorous, national-level dataset from the United States EPA, the US Environmentally Extended Input-Output (EEIO) model, as the foundation for our calculations (learn more about how we do this here). Our updated approach is not only more consistent, it is also more granular: it increases the number of transaction categories in our library from 36 to over 300. This means we are able to more accurately estimate a user’s footprint based on what they buy.
2. Localizing carbon estimates for gas and utilities. The Carbonizer now integrates zip code-level data and supplemental calculations to improve our carbon footprint estimate for gasoline and home utilities purchases. As prices for both gasoline and home utilities vary widely state-by-state, we use your zip code to identify the average price of gasoline or utilities in your area. We then combine this information with your transaction amount in dollars to better estimate the actual quantity of your consumption. With the greenhouse gas intensities of gasoline, natural gas, and electricity provided by the US EPA, we can then estimate the climate impact of your actual usage, rather than relying on national averages.
Looking forward: Room for improvement
We are constantly looking for ways to improve the rigor, relevance, and actionability of our carbon footprint estimates. In this spirit, there are a number of ways in which we anticipate continuing to improve our methodology moving forward.
Further improving the granularity and actionability of estimates. We have identified areas in our category mapping that could be made even more specific to better inform action. For example, some purchases are categorized as “Convenience Store,” which alone, does not provide a clear picture of the types of products purchased. By contextually soliciting more information about more ambiguous purchases, Joro can deliver more nuanced insights and recommendations.
Expansion to global markets. The climate crisis is a shared global challenge. Part of our mission is to help people connect with others across the world to engage in meaningful collective climate action. Our methodology is currently based on US datasets. Moving forward, we aim to integrate macroeconomic datasets from different regions.
Revealing the emissions we cannot reduce alone. Our current methodology does not capture the footprint of publicly-provided goods and services, such as public education, infrastructure, and government services. These services are part of an “immutable” portion of our carbon footprints, which cannot be changed through personal action alone. It is important to acknowledge that this is a meaningful part of our footprints that we can influence through civic action, activism, and engagement. Looking forward, we hope to make this part of our footprints visible and actionable, too.