Editor’s Note: Gavin is a speaker for ODSC East 2022. Be sure to check out his talk, “Timing IoT Devices to Slash Carbon Emissions at Scale,” there!
The largest single driver of climate change is electricity use. So each time we flip on a light switch, charge our phones, or even plug in an electric vehicle, are we helping cause climate change? In 2022, increasingly often, the answer is actually no. These days, conserving energy actually doesn’t always help the environment.
Why not? Each time a new source of electricity demand turns on — say, plugging in an electric vehicle to charge — that extra electricity has to come from somewhere. (A little known fact about the power grid: at any given time, it’s typically a single, “marginal”, power plant that provides that extra energy.)
And, thanks to the tremendous growth of renewable energy in recent years, at many times and places that “marginal” power plant is surplus wind or solar energy. Because the times that people need electricity don’t always nicely match when the wind blows and the sun shines. So using energy at certain times just means that clean power doesn’t have to get thrown out.
But how do we know whether now is one of those times? Or what its emissions will be later? It’s not like power grids share this information publicly.
So, that’s what we do at WattTime: we use a variety of machine learning techniques, particularly in causal inference, to detect and forecast “what times” using energy is cleanest. Every 5 minutes, we scan the world’s power grids and analyze how much pollution using energy right now would cause (if any).
Then we take it one step further, serving those data up in real-time as a REST API that enables IoT devices like smart thermostats, EVs, HVAC systems, batteries, and even industrial loads to automatically shift load to less polluting times. We call this technology Automated Emissions Reduction (AER), and our tools make it possible for manufacturers and operators of smart devices to instantly soak up that surplus renewable energy shrink their carbon footprint.
The reason we can do this in many countries is that we analyze open, free government databases of what power plant emissions, like the US EPA provides. But what do you do when even the power plant emissions themselves are secret?
You monitor the plants yourself. Supported by the Google.org AI Impact Challenge, WattTime and our partners Transition Zero in the UK have begun applying computer vision to near-real-time satellite imagery of power plants to monitor their emissions from space. That project later grew into Climate TRACE, a global coalition of nonprofits and tech companies WattTime cofounded to bring radical transparency to global emissions data. (More here in my recent TED Talk if you’re interested.)
Want to learn more about timing to slash emissions? Come learn more at my session, “Timing IoT Devices to Slash Carbon Emissions at Scale,” at ODSC East on April 19 at 12pm EDT. (Register here and mark your calendar.)
Finally, if you’re interested in learning more about how you can support WattTime’s mission, our team is growing quickly, and we’re regularly hiring for new roles. Join my April 19 session for an introduction to our work, and then check out our open positions and consider applying.
About the Author/ODSC East 2022 Speaker:
While a PhD student in energy econometrics at UC Berkeley, Gavin McCormick invented “Automated Emissions Reduction”: software that instantly reduces pollution from smart devices such as electric vehicles and smart thermostats. Today he is Executive Director of WattTime, a nonprofit that helps Fortune 500 companies and governments use this technology to lower their carbon footprint at scale. He is also a co-founder of Climate TRACE: a coalition of nonprofits, tech companies, and universities using satellites and AI to monitor every source of greenhouse gas emissions source on Earth.