How can data science help in improving carbon management and driving carbon action?
The value of data for businesses
“Data is the new oil.” – This quote has been widely used to highlight the value of data for businesses since the mathematician and entrepreneur Clive Humby first stated it in 2006 (1, 2, 3). Today, data science, the field specialized in unlocking the value of data, is widely adopted in companies and often declared as one of their top priorities (4). However, while companies have heavily adopted data science for driving their business, its potential in managing and reducing their carbon emissions has received comparably little attention so far.
Data science in the area of sustainability
Before going into the specific opportunities for companies, let’s have a quick look at the broader role of data science in the field of climate change and sustainability. Already in 2017, UN Global Pulse hosted a competition on the topic of data for climate action asking for contributions in the areas of climate mitigation, climate adaptation and other sustainable development goals (5). In this competition, researchers and data science practitioners came up with diverse project ideas such as “Predicting and Alleviating Road Flooding for Climate Mitigation” or the award winner “Electro-mobility: Cleaning Mexico City’s Air with Transformational Climate Policies Through Big Data Pattern Analysis in Traffic & Social Mobility”. In the last few years, we also see political organizations like the European Union, nonprofits and big companies investing into projects leveraging data science for sustainability (6). As an example, the Microsoft AI for Earth initiative has been working on assisting ocean and river cleanup by using machine learning for identifying plastic pollution on images.
A more systematic and very thorough assessment of the opportunities of data science in fighting climate change was published in 2019 by Climate Change AI, an organization composed of interested volunteers in academia and industry (7, 8). The propositions cover a very wide range of topics including strategic political decisions (e.g. evaluating policy effects) and traditional environmental protection projects (e.g. reducing deforestation) as well as innovative solutions for businesses. The latter cover different sectors like electricity, logistics, buildings and industrial manufacturing. To give you an idea of the diversity of topics, let me just mention a few of the opportunities identified as high leverage: Forecasting supply and demand for managing electricity systems, freight routing and consolidation in the area of transportation, machine learning driven control of smart buildings, and developing climate friendly chemicals.
The potential of data science in reducing the carbon footprint of companies
Specifically for businesses, BCG recently estimated the carbon reduction potential by using data science and artificial intelligence as 2.6 to 5.3 gigatons of CO₂e until 2030, which amounts to 5% to 10% of the reduction needed to meet the Paris agreement (9). According to their report, data science can be used to improve the whole carbon management process from monitoring to predicting and reducing emissions. At Cozero, we believe in this potential and are making a massive effort to develop data driven features for carbon accounting and decision making, e.g. a carbon accounting accuracy score, forecasting carbon emissions or recommendations of carbon actions. Besides the potential of data science in carbon management – which is what we leverage at Cozero – BCG also highlights that data science enables carbon action by improving efficiency in operations, e.g. in transportation or industrial production.
As examples of such data science projects already helping businesses in reducing their carbon footprint, let’s look at two initiatives from very different companies: the routing optimization software by UPS and the effort for carbon-intelligent computing in Google data centers. UPS has been using ORION (On-Road Integrated Optimization Navigation), a software for data driven routing optimization, for more than a decade now and is constantly improving the algorithm (10, 11). Thereby, they save millions of miles driven, which leads to reducing fuel and labor costs as well as carbon emissions. The ORION project gives a nice example of how data science initiatives which are initially mainly driven by the idea of cutting operational costs can lead to savings in carbon emissions as a co-benefit. Recently, however, we also see some data science projects which are specifically designed to reduce carbon emissions from the beginning, and if necessary achieve a tradeoff between operational costs and carbon emissions. As an example, Google’s carbon-intelligent computing project uses data science to reduce the carbon footprint in their data centers (12). The goal of this project is to carry out as much computing load as possible with low carbon electricity. To achieve this, forecasts of carbon intensities of the electricity grid at different hours of the day and in different regions are leveraged to shift load in space and time, and thereby reduce the carbon footprint.
Back in 2011, Jeff Hammerbacher, the former lead of the data team at Facebook, commented on the widespread effort to use data science for online advertising by saying “The best minds of my generation are thinking about how to make people click ads. That sucks.” (13). Luckily, a lot has changed since then and companies and other organizations have been using data science for an increasingly wide range of topics. Today, we already see the first small contributions of data science for reaching sustainable development goals, and there is potential for much more. Let’s keep following this journey and make a joint effort in leveraging the value of data science for carbon management and carbon action.
Are you interested to learn more about data science at Cozero and how we leverage sustainability data? Schedule a demo here.
Text by Anja Ruisinger, Senior Data Scientist at Cozero