Phenomenon window

Can AI Keep the Lights On? Testing Smart Grid Ideas with Real City Data

You've probably heard that smart grids are the future—AI that can predict energy use and cut carbon emissions. But here's the catch: most of those AI models are trained in perfect simulations, where everything works flawlessly. Real cities are messy. Neighborhoods have different densities, weather changes, and people use energy at different times. So, what happens when you take those smart grid ideas and test them against actual city data? In this project, you'll dig into real-world data to see if the hype holds up. You'll explore how factors like neighborhood density, time of day, and weather affect CO2 emissions across different urban zones. And you'll build simple machine learning models to see if they can predict emissions better than a simple baseline. It's like being a detective for the grid—uncovering patterns and seeing if AI can really help keep the lights on sustainably.

• “

  Think about your own neighborhood: when do you notice the most cars on the road or lights on in houses? How do you think that affects local emissions?

• “

  Have you ever experienced a blackout or a time when the power went out during extreme weather? What do you think caused it, and how might smarter energy management have helped?

• “

  If you could design a smart grid for your city, what one change would you make to reduce emissions, and why do you think it would work?