Polluters Are Paying for Climate Damages. Data Science Is Why
Here are some secret weapons against major corporate carbon emitters
The UK just ruled that fossil fuel companies like Shell and Equinor cannot continue developing projects on the Rosebank oilfield and Jackdaw gas field. This is one of many historical moments in climate-related justice that are happening around the world as we speak.
Governments and courts have become increasingly sensitive to heavily polluting companies. This doesn’t mean that they changed their stance on climate change and pollution. Legal moves like the one by the UK are rather enabled through data science. Being able to prove beyond doubt that polluters are responsible for climate change and for major weather disasters is a key enabler to lawsuits and policy shifts.
Legal momentum is building because of the capabilities of data scientists. Like many things in data science, conducting such analyses is complex and involves a lot of legwork. On the other hand, many good databases and tools have already been built.
This means that even with limited time and resources, a decent data scientist is able to make a meaningful contribution to holding polluting companies accountable. We’ll cover some tools and projects below.
At this point I want to emphasize that I am not in the game of painting fossil fuel companies—or any other high-emitting sectors—as the devil. They provide a meaningful product, energy, and it has allowed us to attain today’s levels of economic prosperity.
However, I do think that heavily polluting companies need to face accountability for the damage they cause. Over time, fines, laws, and lawsuits will motivate them to drastically reduce the side effects of their current way of doing business. Nobody likes pollution and weather disasters, and a powerful way to mitigate them is to have those pay who caused them.
The Challenge: Proving That Polluters Are Responsible
Attribution analyses, where scientists demonstrate beyond doubt that an extreme weather event was caused by environmental pollution, are gaining a strong foothold. However, they remain complex because weather events happen out of a combination of natural and human factors. These can be difficult to untangle.
Also, there are significant uncertainties with today’s climate models. This means that every attribution analysis comes with error bars which can be significant in many cases.
On the legal side, we are only just building a repertoire of precedents. Lacking these precedents earlier meant that companies and courts did not know what to expect from climate lawsuits in the past. Climate is not the only area that faces this situation—think about legislation around AI—but based on precedents it remains a challenge to predict whether a company might face a fine and how big it will be.
Data availability is another common problem that is also present in climate-related data science. There is plenty of satellite data out there to help with weather events, but emission data remains scattered all over the place despite plenty of efforts to centralize and standardize it.
Naturally, the perhaps biggest hurdle in attribution analyses is that you’re creating powerful enemies for yourself. Fossil fuel companies have become more cooperative in recent years, but they have a well-documented track record of denying climate science and blaming everyone except themselves. For an individual data scientist or a small team in a startup like mine, this can feel like a David-versus-Goliath situation.
All this being said, the average data scientist enjoys a good challenge. Then again, you shouldn’t reinvent the wheel either. So, before we look at some cool projects that you might decide to work on, here’s a roundup of resources that one can draw upon.
Data Science Projects Taking on Polluters
Not every project and database is created equal. I see four main categories across demonstrating attribution, assessing economic damages, identifying the biggest polluters, and studying legal precedents. Below are both existing projects but also databases, because all of these will help you get started if you decide to contribute to a new or existing project.
Climate Attribution
The World Weather Attribution is an initiative by climate scientists to understand who caused extreme weather events. They regularly publish scientifically outstanding analyses, most recently on the wildfires in Los Angeles. Reading these can give you a good feeling of what methodologies and skillsets you might need.
The Climate Attribution Database is a massive repository that groups papers and analyses on climate attribution. This helps find precedents to similar weather events happening right now, and it’s a good resource to draw upon in general.
Economic Damages
The NOAA maintains a database on disasters that wiped out more than a billion dollars in the U.S. The billion-dollar database has listed 403 such events to-date. It is a public database and features some in-depth scientific analyses too.
Satellite data service ArcGIS has several tools to assess climate-related damages. Some outstanding ones are the map of economic damage from climate change in the U.S. and the climate global impacts map.
Carbon Emissions Data
Climate TRACE uses satellites and other sensing technologies to spot emissions and tracks down their source. Their data is free and available for download and via API.
The Carbon Disclosure Project (CDP) is a well-established data provider. They’re not free, however, and more geared to larger teams with good resources.
Legal Precedents
The Sabin Center at Columbia University maintains an excellent database of legal precedents. This can be useful for example in finding similar cases and predicting the outcomes of ongoing ones.
The Grantham Center at London School of Economics and Political Science (LSE) regularly creates working papers and policy publications. This can be very interesting again for bridging the gap between data and its legal implications.
What’s Next? Data Science Projects That Still Need to Be Done
Whether you work in the sector or not, whether you can do this at your job or make one of the below ideas a weekend project—we need all hands on deck to help build climate accountability. It might even lead to more professional opportunities down the road for you!
Real-Time Emissions Attribution Model
Climate TRACE has already built capabilities to spot annual emissions and understand which industrial sector is causing them. This is useful for broader accountability but not enough to bring a polluting company to court.
Leveraging existing satellite data and crossing it with corporate disclosures and independent sources might result in a more powerful project. The goal would be to track emissions and attribute them to specific companies in real time.
This is a huge undertaking which would keep droves of data scientists busy for a couple of years at least. However, one could start small by focusing on a specific geographical site where data is of high quality, and then build it from there.
Predictive Litigation Model for Climate Lawsuits
Now that LLMs are eating the world, we might as well train up an instance on climate lawsuits. To support it, one would use case law, precedents, and corporate financials.
This would allow governments, activists, and lawyers to focus on cases with a high potential for wins (and potentially big fines too). Unlike the real-time emissions attribution model, a pilot of this kind of project can be launched even as a single data scientist on a weekend.
