 When it comes to energy regulation, a little bit of data can go a long way. In the United States, we have a wealth of information about our electrical grid. Unfortunately, much of it is not clean, accessible, or connected. The catalyst cooperative collates, analyzes, and prepares data to help advocates close coal power plants. What we'd like to do is share some things that we've learned about how open-source data can further the policy conversation using real examples for our work. My name is Pablo. I've been working with Catalyst as a contractor since December. And this is Christina, who is the president of the Catalyst Cooperative. Christina, can you tell us what you've been up to? Yeah, sure. So a few years ago, a few buddies and I got together with our laptops, scraped some data about our electric utility. We were in Colorado at the time, and we helped convince our utility Excel Energy to close two of its coal plants and then later pledged to move towards 80% renewable by 2030 and 100% by 2050. That's it. We're done here. That's very, very cool. The thing is, if you're going to be a superhero, you actually need to tell the whole backstory. So maybe you can fill us in with a little bit more than that. Sure, fine. Of course, a fight like this one is never fought alone. I'm sure many of you have seen headlines across the country of coal plants closing or the fact that coal is dying and all of these things. And it is, and that those things are happening, but that they're certainly not happening without a fight and they're certainly not happening fast enough for the climate. Our part in all of this started with local advocacy and education. I was looking into some of the remaining coal, remaining capital costs of coal plants, the undepreciated value of those plants, and depreciation is weird and arcane. So I had some questions. So I ended up reaching out to a previous regulator in Colorado named Ron Lair. He was actually setting up a group of folks to come together to try to figure out how we can collectively get ourselves out from under the capital costs of coal so we can transition to a clean energy economy, which was a great timing for us. So it was a nice little gang to fall into. Okay. And when you say get out from under the capital costs, what you're saying, I mean, to me, right, and this is something I think that should come across, the coal power plants don't want to divest from coal until they've actually made back some money on their investment. Is that correct? Yeah, exactly. The utilities need to like, want to be made whole. They can't just like, think it's like walk away from their, their investments. And so we were trying to find targets for early retirement, and we were trying to propose various ways of dealing with those, with that remaining capital costs through refinancing or through other like regulatory mechanisms. And for us, a good target meant that the coal plants, that was a coal plant that was particularly expensive, which is bad for the right pairs, particularly polluting, which is obviously bad for the climate and bad for the local communities. And but those plants didn't have a ton of local, a ton of recent investments, because it's much harder to get out from under a bigger pile of cash. And on this, on this like general team of advocates and financial analysts and data monkeys, unsurprisingly, we were the data monkeys. Okay. And so what exactly did the data monkeys do? Well, so we collected information about these two crucial pieces of costs of the plants, the capital costs, as so they're like all the money that's ever been poured into the plant, as well as the, the operational costs. So, yeah, the operational costs, which also card called the marginal costs. Okay. And when you say the operational costs or marginal costs, essentially what you're describing is how much money it costs just to turn the coal power plants on. Exactly. If you think of, if you think of a plant having a big switch or a big knob, it's how much extra money gets eaten when you turn on the switch or turn up the knob. And so we were trying to, with that, with those operational costs, we were trying to hit those existing operational costs with the all in costs of new wind. That's not a particularly fair fight because it's all the full costs versus just the operational costs. But the utilities have already poured money into these plants. The capital has already been spent. And because of the way they're regulated, they more or less get, they're like, they're more or less guaranteed to get that money back unless they spent it imprudently. And that's a whole of a story. That's another talk. But we were trying to find those plants that could immediately, that if we turn, if we stopped operating them, we could immediately save money for the rate payers by switching to a new renewable source. And we found, in Colorado, we found a few of those unicorn plants, which was very exciting. And what the utilities are, it just because a plant was potentially uneconomic to run, utilities didn't just voluntarily turn them off on their own. They had to be nudged. So armed with that data, the advocates did what they did best when they went and advocated. So they made the case to the various stakeholders at play. So with the utilities, they were like, hey, this is not going to break your bank. It's not going to actually be bad for you. It might actually be good to swap coal for wind. In this case, also with the governors and elected officials, the case there is like, if they care about climate, we can make big moves on climate. Also, they don't care about climate. We can potentially still save rate payers money. And for the regulators, I mean, they were trying to make the case that it was at least plausible that this should be a future that they should look into, because regulators in theory are there to keep rates low. And if it's a possibility that we could save rates by doing a new and potentially better thing, that's their job to help investigate that. So with all of that pressure, Excel, especially from the environmental advocates, Excel came out and said, okay, fine, we'll close two of our coal plants. And once they got sort of familiar with that closure process and got cozy with it, they came out with the pledge to say, hey, we'll reduce our emissions by 80% by 2030 and 100% by 2050. And with the newly minted Colorado legislature and Climate Hawk governor, they like quickly signed that into law. And now it's the law of the land in Colorado. So that's a pretty great result. Yeah, it was a fun process. It was long, but it was fun. And in our part, it showed that we could make a big impact with just a little bit of data in the right hands. And our broader team wanted to do more. We wanted to do this for more than one utility. We wanted to do this for multiple utilities at the same time. We wanted to target utilities based on these operational and capital costs. And we knew that we could not employ, we employ a lot of manual tactics to actually scrape data in Colorado. And we decided to create Cal's cooperative and create the public utility data liberation project to automate this like data monkeying. Pablo, you've been working with us for a while. What is your take on all of this and what we're up to? Okay. Well, as the contractor, the things that I've seen the most is that you've got a whole bunch of different data sets that have a ton of information. So you've got stuff from Energy Information Administration. You've got things from the Federal Energy Regulation Commission and from the EPA, which I hope everybody knows the Environmental Protection Agency. And they're all of them actually making their data available to the public, but they're doing it their own distinct ways. They've got their own formats and they've got their own standards for how well they keep the data. And because of that, it's sort of just really disparate and it's all over the place. And the work that I've been helping you with and that you've been doing in general has been pulling all that stuff together, packaging it up, and so that you can get one big picture of the entire grid and start asking questions about the state of energy costs in the US. Yeah, exactly. Right now, we're archiving, with Pablo's help, we're archiving all of our raw data sources on Zenodo and then pulling them down and running them through a standard ETL process, the extract transform load. And that process generally cleans, gives the data on sets IDs, normalizes the data. And most importantly, and also most difficultly, it connects the data. So there's, you know, a coal plant that is reported in FERC is also reported in EIA, but there's no shared IDs between the two of them. And we need those in order to kind of ask bigger questions quickly. So we do some of that inner dataset connecting and intro dataset connecting sometimes, unfortunately, one of the best. And then we archive that like cleaned up data in Zenodo data, sorry, again, we archive it on Zenodo, but in the format is Frictionalist Data Packages, which has been fun to work with. And then there's that like on top of that, we kind of pipe it into other data formats and have some like stock analysis that's kind of built in. Cool. So this is the part where we get to do a demo. So we can show these, so we can show folks what kind of question funnel actually answers. Ready? Exciting. Yeah. All right. Can you, I'm gonna... Nice. Great. Thank you. Can you all see this notebook? Yes, we can. Grand. Okay. So I'm gonna go through this quickly, but so there's, this is a little notebook that is exploring these marginal and capital costs of coal plants. There's a ton of prep that kind of just pulls everything into one place. My computer is being slow. But generally, that prep above generates these two separate data frames. One is about renewable costs and the other is the data frame here that pulls in data from FERC, the one dataset FERC and the other EIA about the capital and marginal costs. And we have data in this for 43 states, not all of the utilities and all of the states need to report. So we're not at everything, but the best way to look at most things is plots. So let's play with some plots real quick. So I'm gonna show y'all some information about Colorado in particular, but we can get some crowdsourced states in a minute if y'all want. So this is just a snapshot of the annual operational costs in Colorado in 2018. The purple at the bottom is fuel costs and the turquoise at the top is non-fuel operational costs, everything else that all the other money that's spent when you turn that plant on. And these numbers down here is how big the plants are. It's the capacity of the plants, the biggest ones are on the left. And the way that we kind of compare and sort of do kind of some ground-true thing of whether or not these plants are potential targets is doing that, pitting them against renewables. So I'm going to turn on the renewables in here, so it's a little small. So this is a line of the average U.S. cost of wind and the average U.S. cost of solar in 2018. Solar and wind prices vary quite dramatically across the country. So if you were to pick up this data to run with it in one of your states, there's some local variability and we need to get some more granular data. But as you can see in Colorado in 2018, really all of the plants are more expensive than wind and most of them are more expensive than coal. Now that was very much not the case 10 years ago and we can show that in a few separate ways. More expensive than solar, right? Yeah, so solar is the, right now it's slightly more expensive resource than wind. But yeah, in 2009 this, the landscape looked very different in terms of the economics of wind versus solar. We can also look at these operational costs over time. This is just one of the plants in Colorado and we can look at it against with the changing costs of renewables. So for this particular plant it was around 2017 when it actually became quote-unquote economic to immediately switch. There's a lot of other reasons why a plant is on, like we need to keep the life on, we need to keep the grid running. But the data that we've been trying to compile in these is really to find out like, is it plausible? Could we turn these off? Please utility, if it's plausible that it's economically viable to switch to renewables, we should do additional modeling. This is really cool. I think people can tell we rehearsed some of this, but the downward trend of the costs for the solar and the wind is, that's new, that's actually new to me right now. Yeah, it's been quite fun to see it over time. We can also look at the capital costs and how capital costs are changed over time in plants. This is a particular plant in Colorado which had a big investment in 2014. They got scrubbers installed, which is great for the emissions of that, like that some of the emissions of the plant like socks and knocks, but it actually makes this plant really hard to kind of make the case to retire. So I think we only have a couple more minutes. So if you have any final points or you want to take a question? Yeah, let's take questions. Looks like Zane's been getting some of them. Yeah. Thanks, Zane. That was really great. One I like is, are there ways to get help from volunteers or crowdsourcers? Ooh, yes. Yeah, luckily right now we have several like open source contributors to our project. Most of them are academics, which has been really lovely. It's been great to see folks kind of use, pick up and use our data in their research as well as like add data sets and add to the project in general. So it's been very helpful. We have a ton of sort of pain points that we're always trying to work on and trying to get better about. Yeah. If someone's in the coding side of things, particularly with Python, it would be handy for us to get a boost in. There's a library we depend on for it called dbfread that works in, it basically opens up Foxpro databases because when we said some of these folks are using really old and weird data publication methods, this is one of them. And that needs some attention. And we're interested in taking on the task of making sure that it gets and stays updated, but it would be really nice if we could just play more of a coordinating role rather than having to like individually write all that code ourselves. We're going to try to do it anyway, but if someone's get in that area, we could just have some extra attention. Yeah, definitely. Well, thank you both. We have to move on to the next talk, but thank you very much.