 Stanford University. Next session, we are going to hear about different data we have been collected during this energy transition. And how this data can help us to make a very important decision, whatever is regulatory decision or business decisions. This session will lead by Commissioner Andrew McAlster. And without further ado, I will have Commissioner McAlster have the stage. All right, everyone. Well, thanks for sticking it out for the whole afternoon. If everyone could take their seats, we have a really great final wrap-up panel here. And my name is Andrew McAlster. I'm a commissioner at the California Energy Commission. And we are just really happy to have worked on this and co-sponsored with Stanford just really timely and very meaty content today, as you all know. And through the course of the day, you've heard a lot about data. I think every panel has mentioned, oh, data. We need better access to data. Or data is key for this. Or we're doing time-dependent data. We're doing load assessments. All of the load flexibility, that's time and location. That's data. So the way I think about data is the electric sector. Electric system is the backbone of our clean energy transition. Our programs and policies, innovations, development of all the stuff that hangs from the poles and wires and all the innovations on the technical front, those are the organs, the physical meat on the bones of the actual system. Data is the hemoglobin of that system. Data lets everything else work together. Data is the conductor. It allows the conductor to make sure that literally billions of devices connected to the grid on the supply side, all the way down to the smallest of loads, can play to the same tune, that that orchestra sounds good, that the lights stay on, that reliability is the thing. So data, it means a lot of different things to different people. There are lots of different types of data. Interval meter data, which we're going to talk about some, but lots of different information, lots of different types of information that have to be integrated to create the knowledge that lets us operate the electric grid, the clean electric grid. Really what we're talking about is industrial policy. This is the investments and the level of investments that are coming to California are of the scale of going from biomass to coal, going from coal to petroleum around World War II. There was this massive shift in our economies, the industrialization of our economies and then the shift to more energy dense fuels. The amount of investment that transforms societies in those contexts is really what we're going through now. And we're moving to fundamentally different and cleaner sources of energy. And we need a commensurate, we are going to see a commensurate wave of investment. That's real dollars. If we think about buildings, to get our existing buildings, as Sally this morning mentioned, the existing buildings as being sort of really the focus and I couldn't agree more. At a minimum, we're gonna need $300 billion to transform our existing buildings in California I'm talking about. Transportation's maybe double that if we think of all the EVs and all the infrastructure with those EVs. So we're talking literally trillions of dollars. If we think about the electric grid itself and the equipment that has to be sort of improved and rebuilt over the next couple of decades, we're talking trillions of dollars. And that is each one of those investments, whether it's in sort of traditional supply infrastructure, poles and wires, end use, load flexibility, new types of generation storage. Each one of those investments is an opportunity to make sure that that device is, comports itself well on the grid. It's a good citizen of the grid. So it's a huge opportunity. And so at the Energy Commission and across all the state agencies you heard from Leanne and Alice, we are working really hard to make sure, to create the conditions and sort of regulatory infrastructure so that those decisions have the right signals and the right incentives to play well on the grid. So the investment opportunity, just I can't overstate that. The vast majority of this capital that's gonna transform our electric grid is private capital. We're not gonna subsidize our way out of this. So at the Energy Commission, we're a regulatory agency and we're doing a bunch of things like flexible demand appliance standards, load management standards, lots of technical solutions that really we are the only, that state kind of has to say, okay, we're gonna standardize this and require it. Then everybody has a level playing field and can invest in the sort of IP that's really gonna let this succeed and figure out the business models that are gonna let it succeed. So there's policy development, planning, regulation, as I said. And then everyone, I think we're all in agreement that we're in sort of at the great implementation period. So lots of activities and data really enables all these activities and their optimization. So on the one hand we have a state, the sort of at the Energy Commission, we now have data from every customer in the state and we can use it to do planning, interval meter data, we can look at seasonal loads and do all sorts of grid analysis. But there's also that out there in the world data like Green Button Connect and those of you who say third parties who work with customers to figure out how to optimize their energy consumption on this grid of the future. So again, data's a very, it's a topic that means a lot of different things to different people. We have a great panel that complement each other really well and each speaker is gonna talk about one aspect, what they're using data for, some of their vision about how data can improve, the way we plan and implement the clean energy transition and emphasis on the electric grid. And so these use cases and these visions and these activities that each speaker will talk about, I think will really shed light on sort of this large landscape that for which data is critical. And then we can go into some Q and A and just help elucidate any details that people might find interesting and we'll go to some audience questions towards the end. So I will just introduce quickly or just really let you know who's gonna speak. So Max Alfhammer is UC Berkeley. Well, you can read it there. We've got a Port of Bargava from Weave Grid, Arvigua from Google, and Amber Mahome from E3, and then Mallory Nobles from Stanford. So all these speakers in their own way are helping utilize data. We're fully in the digital age. The monopoly regulated electric sector is the only, it's a little bit behind the game in terms of using data. We're all being tracked in real time by our phones and getting targeted by ads in real time, I mean by the second. So hourly interval meter data doesn't seem like that big of a deal in that context, but here we are. And it's about time that we move this industry and the grid edge particularly into the digital age. And so we're gonna hear a lot about how that is happening and how it could happen. And I think it'll be really stimulating. So we'll start with Max. All right, thank you. Very much, my name is Max Alfhammer. I'm on the faculty at the other large university up in the Bay Area here called UC Berkeley. I've seen a lot of people with business models on the stage today. My business model is essentially to talk regulators and firms out of data they don't wanna give up. So that's the big return on investment here. We were also told we have four minutes each, so I will dive right into it. But one main point I wanna make, data are of course a key necessary ingredient to what we do, but there's just as much innovation in the algorithms we use to analyze these data and the machines that these algorithms run on. So there's just been a revolution in what we can do. And I figured I'd bring you two ideas for my four minutes. Idea number one is we keep on talking about meeting electricity demand, but the world is going to get warmer. That means summers are going to be hotter, people are gonna air condition more, but winters are going to be milder. In Northern California, most of your energy is consumed in the winter, heating your home with currently natural gas. So an economist likes nothing better than a trade-off, so we figured what happens in the summer versus what happens in the winter. About 15 years ago or so, we worked with the California Public Utilities Commission to figure out could we get access to billing data. So essentially what is lots of people have now, we have two billion electricity bills, use statistics and then fancy computers to figure out how do people respond to hotter and cooler temperatures in their energy demand. So this slide on the right shows you how somebody in Pasadena reacts to a really hot day, that's the red line, San Francisco is the green line. The San Franciscans here go what hot days, right? The question of course becomes what if San Francisco turns into something like Fresno in the summers, do people all of a sudden have air conditioners? So instead of just using three cities where you talk to your local utility, we use all the investor-owned utilities data so we can estimate these response functions, how people react, how people's electricity and gas consumption reacts to temperature for every zip code in California that is covered by this data set. So here you're seeing 1,500 different, basically dose response functions of how electricity and gas consumption reacts to hotter days. The big takeaway message here is, yes it's gonna be hotter in the summer, we need all this capacity that one of the commissioners was talking about earlier in the summer that sits idle much of the year to heat electricity demand on the hottest day, but gas demand is going to go down pretty drastically over time because you're not gonna heat as much in the winter, especially in better insulated homes. What can we do with this? Well, what I showed you is interesting to a nerd like me, but we could actually project what happens to electricity and natural gas demand going forward at the zip code level. There's uncertainty around this, but then we can see who is going to see the biggest increases in their electricity and biggest decreases in their natural gas demand. So I'm just also going to make a plug. When you have these panels, yes, we faculty are more likely to own the nice suit to represent on the stage, but all the good work here is done by armies of students in postdocs. So Tamar Carlton, who's a professor at UC Santa Barbara, generated these photos. You should check out her work, she's amazing. I'm just gonna give you a little preview of a study that's coming out in the next couple of weeks. We care a lot about energy consumption data and what happens to demand, but the energy revolution is going to affect so many different aspects of life and so many different areas of policies. So where we put windmills is of course at the front of everybody's mind. Do people hate windmills? Are they ugly or are they beautiful? So open your LA times, right? Are they ugly or are they beautiful? So we have a study where for every housing transaction in North America, we figured out whether the home can actually see the windmill past natural obstructions or not and I'm gonna compare two neighbors for all homes sold in America over 24 years as to who could see it and who couldn't. I'm not trying to self promote here. What I'm saying is this would not have been possible 15 years ago with the computing power and the algorithms we had, this is possible now. So this is not data I talked my way into. This is computer scientists you talk to who tell you which algorithms to use. If you wanna know the answer to this question, you'll read about it in the news in three weeks. I hope if not, feel free to shoot me an email. There's a QR code to my website and off we go to the next one. Thank you. Great. All right, let's see. That was great. Yeah, you and I both have. Great. Thank you. All right, it is the three o'clock session which means that a lot of you are probably a little sleepy. So we're gonna start off in a slightly different way with my session. My name is Empor Bhargava. I'm an alum of Stanford. I started my company out of here. My co-founder and I were both at Stanford. It was the classic love story of PhD meets MBA. One of us wanted to make money. The other one wanted to just further his research. So here we are. But I wanna start by saying something. So I'm gonna count to three and I'd like everyone to clap their hands. It's gonna wake us all up. One, two, three. All right, I saw a bunch of you not clapping. I saw a bunch of you going off the beat. What I just did there was provide you with data on when to clap. There are 280 million vehicles in the United States. There's 120 million households. We're gonna take 280 million vehicles and make them go all electric. And as that happens, they better damn clap at the same time or they better damn clap at the right time. And that question right there is one that is foundationally foundationally impossible to solve without data. If we don't know how people charge, how they might discharge, how they drive and what their driving behavior is going to mean relative to their energy consumption, it becomes incredibly difficult to solve the question of how we orchestrate the way that all of your clapping orchestrated, even those of you who didn't clap, all of that orchestration requires a tremendous amount of data. So we can talk all about the smarts. I'm gonna charge, I'm gonna discharge, I'm gonna do all of those things. But in a vacuum, if each of you clapped whenever the hell you felt like it, it'd sound like cacophony, right? That's why data is so foundationally important. And what my company does at Weave Grid is we build the platform that enables these tremendous volumes of data and controls, because controls is the next layer up, to be orchestrated by bringing together the hundreds of thousands, millions, eventually hundreds of millions of vehicles together with the customers, the automakers, and of course the electric utilities that own and operate the electric grid in a way that this all works together seamlessly. And at the core of it all is a really powerful, yes we call it artificial intelligence today, we used to call it machine learning when I was in graduate school, we used to call it statistics probably a few years before that, but the point is at the core is a really powerful engine that enables all of that data combined with the incredibly disparate controls layers that exist between Fords and GMs and Teslas and Charge Points to work in a singular orchestrated fashion so that we can deliver a more affordable, reliable, and cleaner grid. And I did not use any slides so far which makes me really happy because I threw these together last minute. Very, very quickly though, that's what it is, right? We're basically building this platform, gives you a very simple sense. I think the one thing I wanna leave everyone with though and I promised that I would not speak too long on my slides itself, the core flexibility comes down to not can I manage an EV, can I dispatch an EV, but what am I doing it for? And that challenge is one that we've actually not answered very well even though the data are staring us in the face back to data. The data suggests that the most painful part of the electric vehicle transition is actually going to be on the distribution network. If you talk to Jason or Steve or any of the folks in the first panel, they will tell you it is that system that is most at risk if we do not start managing charging appropriately for that cost. And what most optimization has historically done since 1973, since the energy crisis is focus on the commodity five to 10 events a year that drive up bulk system peak, not the distribution system. And so that's a lot of what we focus on at Weave Grid. I'm gonna skim through the rest of these but really it's about bringing down that distribution cost as well and ensuring that we can optimize the whole stack of value, generation, transmission and distribution. So I'll leave you with that and we're gonna answer some more questions later. So you can fire up the slides. There's this next slide situation. There you go. Hi, my name is Guha. I work on the data commons project at Google. So to give you a bit of context, there's a lot of data out there and I'm talking about public data, data of the sort that the census, the energy information agency, IPCC, BLS and so on collect. This data is essential for everything, for journalists, for policy folks, for scientists. Unfortunately, using this data is incredibly painful. It comes in a wide variety of formats and schemas and it involves expensive data wrangling that is repeated over and over again. Let me give you a concrete example. We know from IPCC data and I'll tell you how we know that sometime in the next 10 years, peak temperatures in your county and Imperial County are likely to cross 50 degrees centigrade with a 5% probability. When that happens, it's gonna be all hell, it's gonna break loose in terms of health, shelter, food and so on and we can't start preparing for that when that happens. We need to start preparing for it now. And in order to, somewhere in my slides got mixed up, but let's see if I can do this. Basically, in order to prepare for that, we need data not just about the climate, but we need data about health, we need data about energy, we need data about housing and these all come from so many different organizations. It's beyond the scope of somebody sitting in one of these counties to put them all together. It's probably beyond the scope of somebody even sitting in Sacramento to do it and it's very difficult even in Washington from an organizational perspective to do it. In order to make it easy to use, somebody needs to organize all of this data and make it easily accessible, which is incidentally Google's mission. So we've been doing this at Google for the last several years, but because this is so important, because this data is essential for meeting the grand challenges of the century, we're doing it in an open fashion. Open meaning the data is open. Everybody, universities, NGOs, governments, even a competitor should be able to build on top of this. Open meaning all the source code, the entire software stack on which this is built is open. And most importantly, open meaning everybody can participate in building this. There's no magic silver bullet in building this. Our approach is to do the data wrangling once and for all and make the data easily available. So we've taken data from many, many, many different organizations and including, okay, somewhere my slides got. From all of these about demographics, economics, housing, there's a huge amount of data about the climate. IPCC data is a great example of open data, but good luck downloading the 100,000 plus net CDF files which are on different grid levels from about 80 different modeling agencies and making sense of it. So we do all of that. It's one of the biggest databases out there, 250 billion plus data points. If you're familiar with Fred, which is a federal reserves economic databases, is about more than five times the size of that. And then we make all of this available to everybody who wants it. You're welcome to download it at several petabytes, but many people don't have that much storage just lying around. So we provide APIs on top of which visualization tools, natural language tools, et cetera, et cetera can be built. Of course, this is all done not just by us, but we work together with many, many different partners including, of course, I have to mention the door school and which brings me back to the example that I mentioned right at the beginning. You know, how do we know this? You won't find this data about these temperatures crossing these thresholds anywhere in IPCC or NASA data. What those, the data sets will tell you are sort of simulations from many different organizations. All of that stuff needs to be brought together and we need to be able to make predictions like this. And this was actually work that was done together with people at the Sustainability School, Aditi and Ramraj Gopal and Arun. So we're looking for partners. We've got lots of partners including people like the UN and Feeding America and others. We'd love to have more. Hi everyone, my name is Amber Mahon. I'm a managing partner at E3 which stands for Energy and Environmental Economics. We are a consulting firm. I'm based up in San Francisco. And we have the pleasure of working with the California Energy Commission on lots of really interesting policy problems, everything ranging from modeling California's renewable portfolio standard goals and SB 100, how to get the grid to zero carbon all the way down to what is the avoided cost of electricity in our avoided cost calculator that informs the state's building code standards. You heard from Leanne Randolph earlier today. We supported the Air Resources Board in modeling their scoping plan, looking at economy-wide decarbonization. We've been working with the Public Utilities Commission on their integrated resource planning, modeling the grid of the future. So there's just a whole range of data analytics problems and all of those public policy questions. And having the ability to have access to this data, manipulate the data is just essential to informing good public policy. And that's really what my goal has been working at E3. My background is actually in public policy, not in data science or engineering. But it's a very interdisciplinary field. So I thought today I'd show a few examples of some work that we've been doing around distribution planning. So, of course, we've talked about the high cost of electric distribution. We know that California electric rates are in part driven by the high cost of the distribution system, including meeting the challenges of wildfire, grid hardening, undergrounding the distribution system. But we also know that if we want to decarbonize the state, we need to electrify all of our buildings and all of our cars. And so we need to be able to interconnect all of those new loads in a way that's equitable, reliable, and affordable. So that means we need to be able to do it in a smarter way. And what's interesting about the electric distribution system is that in many cases, this is a 70-year-old, 8-year-old, 100-year-old system that we don't actually have great data on in terms of the location of some of the distribution poles, the size of these transformers. You know, I've talked to some of my friends at utilities that find that when they send their linemen out, you know, the location of some of their infrastructure is actually not even in the place where they thought it was because the maps were hand drawn 70 years ago. So there's just a lot of effort that we need to do to modernize our electric grid for this low carbon future that we're all trying to plan and build together. And that is gonna require better analytics, but also kind of meshing the old with the new. So jumping in, what is distribution planning in this low carbon future? You know, part of what it means is building new geospatial data sets that layer different types of data together. And that really requires a lot of data wrangling to the earlier point. And then also some projections around, where do we think these new loads are going to appear? What do we know about customers and can that help us inform predictions around where customer adoption is going to take place? Max made a great point around as the climate warms, what are the tipping points for air conditioner adoption? That's a non-linear trend that we don't have a lot of historical data to look back on to inform. But we need to be able to project that if we're going to be able to smartly plan the distribution system. So this is just an example of one of the things that we're trying to do at E3 that is building sort of a geospatial planning tool for the electric distribution system. What does this look like? This is just an example for EV charging. One of the things that we're trying to do is help understand where to build EV charging stations. And that's partly informed by driving patterns and customer demographics, but also the existing extent of the electric distribution system. So there's a big spatial element to this data challenge, but there's also a temporal element to that challenge. So we know that hourly electricity demand drives our planning paradigm. California today is largely a summer peaking grid, but interestingly, that's actually not true in different parts of the state. So we know that in the East Bay of California and Oakland, there's not a lot of air conditioning loads, but there is electric resistance space heating. And so it's actually a winter peaking system in the little portion of East Bay there. When we look at other parts of the country that are very cold, this is an example from some work that we did in New York City. We see electrification of buildings driving a winter peaking system in that region over time as you move to electric space heating from heat pumps. So that creates a whole new planning paradigm for the distribution system when you've gone from planning around meeting your summer peak to suddenly now you have a dual peaking or a winter peaking system. So really what I just wanted to illustrate are a few interesting charts that this, the questions around planning for the grid of the future involve temporal questions, it involves spatial questions, it involves layering different types of data sets that we haven't thought about before. There was a lot of talk about equity earlier and we have great data in California around the location of disadvantaged communities. You can go on, I think, on the CSUS website and pull down these amazing maps with a lot of demographic information but then the question becomes, how do you value these different characteristics? If we're trying to plan the distribution system in a way that's equitable, how do we value the equity, the resilience, the cost? It becomes not just a data analytics question but a public policy and prioritization question as well. So with that, I'm happy to turn it over to Mallory. Great. Hi, everyone. Hi, everyone, I'm Mallory Nobles and it's great to be here today. I'm here to talk to you about one of the initiatives that Stanford is using to train the next generation of data scientists to make progress on important social problems and this is through a new major that we've launched called the Data Science and Social Systems major and this launched last fall and our goal through the major is to equip students with three fluencies. We want them to become experts in statistical and computational methods to have a set of core knowledge around the social sciences and to also have deep and interdisciplinary expertise into a particular social problem. And we find that a lot of our students are very passionate and interested in working climate and the environment and energy issues. And so with these three sets of fluencies, we train students to address social problems using data through a framework where in the first step they frame an important problem. We ask them to consider what does the academic research tell us about this problem? What do the lived experiences of those who are closest to the problem or those who are working on the problem tell us about this issue? And also what can we learn from data? Then we give them a set of skills so that they can design research methodologies and implement these research methodologies in machine learning and statistics and optimization and in causal inference to be able to create research, to inform policies about these problems and then finally remind them that once they have the results from a model that they are confident in, this is really the start of using data to make progress on an important social problem, not the finish line. That there's a lot of work done between translating the technical or academic knowledge into policy and real change. And towards the end of their time at Stanford, they participate in a capstone where students work in small groups with an external organization, either government, industry, or nonprofits. And we're really excited that the CEC is one of our pilot partners for the capstone. We have a team working with Erica Brand, Safia Hazenday and Emily Leslie to continue their work on determining the optimal location for wind and solar resources, taking into account land use, natural resources and social factors. And last fall, I was teaching our gateway course to the major to a group of 80 students and Emily was gracious enough to come and share some of the work that she has done in this area with our students. And we had a really positive response to this work. So two things that stood out to me from what students' responses were, were A, that it was really cool to see a social problem where there was such a tangible feeling of how data science was being used to make progress. I think the timeline that you all are working on doesn't leave the luxury of necessarily making small or theoretical progress that is important to really make tangible steps. And that they were also excited about that this area and the CEC was able to build relationships so that the work that the data scientists were doing was able to translate into real policy. And so when I think about, I see students that are really excited about learning a lot of technical skills. I see students that are really passionate about climate, energy, the environment. And I think the piece that's left for us to make progress on is how can we make it clear to students what are the opportunities in this area. And so I was really excited to see the question in the last panel about addressing the elephant in the room and how do we support the next generation of students. So my call to all of you is to think about within your own organization how can you develop research opportunities, internships, entry level positions. And then also how can you work with other organizations in this room and in this space to think about how can we make it clear what a career path looks like for a student who's interested in working in this space. I think we can learn a lot from Big Tech in this area. They've made it very clear what the steps are to become a successful software engineer. So how can we do the same when thinking about if you're a person who's passionate about using data, cares about climate and the environment and energy, what does that pathway look like for you? And so I have a plug if you wanna connect with me but also encourage you to also think about connecting with all universities and the younger generation. Well, thanks all of you, super instructive and I think it warranted a little more time than what we had asked you to present in. So I wanted to actually build on something on Mallory's presentation first. There's urgency. We have to get moving here. We don't have the luxury. I mean, so it's who are the folks that are gonna be putting together these systems and automation and algorithms to really optimize and target investment and figure out how to do load shaping and all that stuff. So we really need sort of not only to train young folks but also to take folks who are already at our institutions and sort of give them the tools that they need to get better on data, right? And I wanted to just ask each of you, I mean, we have academia, consulting, startup, private sector, sort of all doing different things with data. Could you just sort of describe what it takes to bring in and train up a data sort of member of your team to be useful in handling large data sets and doing these kinds of analyses? Sort of what does that look like in practical terms? Maybe we'll just start with Max and go. Start on this end and go that way. And then next one we'll do the other way. Yeah, I mean, usually what I tell undergrad students they come in first day with their 4.6 high school GPA and they say, I wanna change the world. I was like, let's go do it. Take some statistics, learn how to code in R or Python or Julia, whatever it is you're going to do. And then I think one point you made that is so important. Figure out what framework is appealing to you if you're an economist, an engineer, a biologist, whatever it is you're interested in but become trained up in sort of an intermediate level knowledge of what a core major would understand and then start engaging. So this notion of you can just dive in today, it doesn't take any foundational skills as a fantasy. So in Econ, that's math. The New York Times proclaimed that we're all climate economists now. Even people in finance are becoming interested in climate change, the macro economists. So new tools are being developed but I'm, stats and computation are sort of the building blocks in what I do. Well, gosh at E3 we hire a lot of really smart people that come to us with their great GPAs and also some work experience often. And we find that we still need to train them. And so we have a pretty intensive, it's almost like a Stanford course. In fact, sort of modeled on a course that's taught here at Stanford to teach people the fundamentals of the electricity sector and regulatory economics to kind of bring everybody up to the same language so we can all kind of speak the same language. And then there's a lot of on-the-job training and we find that working with clients and being curious and asking lots of questions is the best way to learn and grow. Great. I'm gonna piggyback off something that Amber just said of learning to speak the same language. So I think we want folks who identify as a technologist to be able to communicate with policymakers. We want folks who identify as interested in working in policy to be able to communicate with the technologists. And so making sure that you have a broad set of skills so that you are able to be that connector and translator is also important. Great. We get a lot of smart people too, I think, at Google. They come trained, et cetera, and we do all kinds of things, but that's not what I wanna pitch. I wanna talk about, yes, it's important to know statistics and programming and so on and so forth, but we believe that data literacy is need not be restricted to people who can write code. Every person out there needs to know what a bias sample is, what so many of the basic concepts are. So one of the things that we've been doing is we've developed a course, which we call Data Science for Non-Programmers, and we are actually piloting it last quarter, last semester, we piloted it at Claflin University, which is one of the HPCUs, it's in North Carolina, and we're rolling it out to a larger number of people over the next few semesters. We're also piloting it in some places in India. I would generally believe that in a data scientist, we just assume that everybody knows how to use a calculator, a phone, and things like that, and people need to be literate about these things. We have a slightly easier hack. We just hire them from Google. We let them train them so we can hire them. Options, the power of options. Yeah, exactly, power of options. I mean, jokes aside, right? Like as a startup, myself, my co-founder, a lot of our team came from deep expertise in respective parts of the industry that we work in, and our industry is not a singular one. I have people who've worked on the best amount of response companies known to mankind historically. Guess what? This is a whole new adventure for all of them. The Nests and the Internox of the world, the O-Powers of the world, that's where a lot of my utility-facing folks came from, it is a new adventure for them as well. We've got a ton of people from the automotive industry, GM, Tesla, Ford, you name it, we've got it on our team. All of them are dealing with a whole new set of problems. At the core of it all, then you've got data science, right? And you've got to figure out how do you build the right data tools that, by the way, because we deal with grid edge load flexibility, we're not just talking about analytics on the edge, we're talking about controls on the edge. How do you optimize? So when you send a signal, and it just fails, because the 4G network failed, how do you make sure somebody doesn't wake up with 10% state of charge when they expected it to have 100% state of charge? There are failure modes you have to build in that aren't the same as whether your Snapchat works or not. And that is a very, very different frame of reference. And so it actually all starts with being able to speak the same language, as you said, but also having shared empathy. And that's actually one of our core values of the company, empathy for the customer, empathy for the utility, believe it or not, empathy for the automaker, the partners that we deal with, the policy makers that we interact with. Empathy for the fact that this is a hard job, we're trying to transform two massive sectors in the span of the next 25 years. It's not gonna happen easily, and none of us are gonna have an easy time, right? So you have to have empathy for that in order to solve the problem. Which then allows you to ask yourself, what questions do I need to resolve? What answers do I need to deliver? Which allows you to then dig into what data do I need? That then allows you to go develop those solutions. Well, beautiful, and I couldn't agree more. At the end of the day, we all talk about tech, and many of us are engineers, economists, and stuff, but these are all human decisions and human problems, and I think it really has to be human-centered. As we, yes, it's data, and we're in a cubicle on a computer, but really these are fundamentally our decisions in this transition affect people in just a very, very real and tangible way. So I think we have to center our decision-making on that, so I could not agree more. And we are having a similar kind of experience at the Energy Commission as we create this data lake and just have access to unprecedented consumption data, electric and gas, and related data for forecasting. And we have about 750 people, and we have a site license now for Tableau, and are training people up, and people who are mid-career or later are now really, I think, inspired to actually touch this data, figure it out, talk to somebody who knows more than them, and really come up with ideas for what questions they might now actually be able to approach, right, which they haven't before. One of those, say in our world, is okay, how do we get people to, when their AC goes bad, do they put in a heat pump? How do we understand the impact of that? That has all sorts of rates, locational, climate, behavioral, income, all sorts of facets to it. And now we can approach that question when we really couldn't before. And I wanted to just ask, maybe starting with Goua, how, what are sort of, how do you get innovation, like I said, the data commons, you're integrating all this data. I always think about how do we try to monetize, say, healthcare outcomes that reduce healthcare costs into pay for induction cooktops, right? So those sorts of questions, I'm wondering, just across all, not even related to energy necessarily, but across all the various sort of data sources that you're integrating, what are, you know, what's an example or two of just the really interesting questions that you now, that sort of, you didn't anticipate that people are working on. So one of the things that we learned, being in the middle of Silicon Valley, being in tech, we think we have all the answers. We realize that the thing that comes out of these, of this data, is stories. And the people who know the stories, who can tell the stories of the people who know about the content. So recently, there's an NGO called One.org, which is Bono's NGO, and they did this, they took a whole bunch of data from us, and they did this deep study of where is all the money that has been given by the developed countries to the developing countries for climate mitigation going to? And it turns out that of the 800 billion or so that had been claimed to have been distributed, less than 400 billion has actually gone out. And less than 100 billion has actually gone to people, gone for these things. Some of the money has actually been used for building things like chocolate factories. I'm a big fan of chocolate, but it's not quite in climate change. Another really interesting thing that came out, that a bunch of healthcare people is, they were, you know, sometime late 2020, they were trying to build a predictive model to predict COVID morbidity, which is what fraction of the people who get COVID actually fatalities. And in a typical study like that, you take, you know, 10, 15, 20 covariates. Over here, there are 60,000 covariates, threw it all together, built the cause of model. And one of the fascinating thing is that one of the best predictors of COVID morbidity is the number of prescriptions per capita, as per the data release for the Drug Enforcement Agency. In the US, when you fulfill a prescription, that it could be for blood pressure, it could be for diabetes, whatever, it goes to the DEA, which then aggregates it and republishes it. If you step back and think about it, it makes total sense because that is one of the best indicators for pre-existing conditions. And you're just like, what? And it's also something that you know for every county out there. These are the kind of things that come, but in order to get those stories out, you need to go talk to the people who know the content. Now, often these are not computer scientists who don't have the ability, who can't afford to spend the resources to get all the data together and do all of that stuff, but if they can make things much, much, much easier for them, then these stories can come out. Beautiful. Does anybody else have an example that just kind of blew them away, that now they're able to approach, that they weren't until pretty recently? I mean, I think one of the areas tangentially energy related, but Marshall Burke here at the Doors School is looking at outdoor indoor pollution monitors, what happens in a wildfire event, how much pollution goes into the house, and what does that mean for health outcomes? So just our ability to combine what goes on in the energy economy, and I'm not gonna blame utilities for sparking wildfires, but there's a correlation there, right? And what that means again for human welfare and what that then informs in terms of the policy choices we make here is I think really, really interesting. One more example, this is actually work being done by Ram Raj Gopal and one of his students over here at Stanford, working together with Data Commons. I don't think it's been published yet, maybe Aaron can tell us whether it's been published. It's about the correlation between income categories and in various metrics of social disparity and power outages. It's really depressing to look at it, but you have to tell the story for us to recognize that this is happening. I think that paper's coming out sometime later. That's beautiful. I mean I think one element of climate change that I think is unique is that none of us will get to our goal on climate change unless everybody gets to our goal. Like it's the fundamentally common property problem, right, so there really is no, climate doesn't discriminate. So I think all of us feel the urgency and everybody has to come along. So we're gonna have to solve some social problems that we somehow haven't solved before, but we're gonna have to solve them if we're gonna get to the finish line together. So I really appreciate the tools that are gonna enable us to really, really sort of hold ourselves to account. Let's see, I guess what, so this is a self-interested question, but we've heard a lot about data access and what a hassle it is and what the barriers to data access and how the cost and cleaning up data repeatedly and just all these kinds of barriers. What, is there anything that the state can do to sort of facilitate, outline sort of the data that the state is collecting and sort of working with ourselves, but also sort of data access for service providers who can help consumers who really need help to understand how they can help, sort of the data access more broadly. Where are you seeing sort of barriers and what might you think the solutions are? Maybe let's just start here and go across. Yeah, sure thing. Well, as we learned last night, the 49ers did beat the Lions and why that's important is we're not just talking about access to energy data, there is now increasingly a massive co-dependency on transportation data and that's very difficult to access. It is very hard to get the increasingly huge volumes of telemetry that automakers are using constantly for card credits, for example, LCFS credits or other things like that and even if you can access it, my company happens to be one that works very closely with the automakers to be able to build those data pipelines. It's really hard to create it in a standard way and I think importantly, to get it to the specification required to build confidence in achieving the outcomes that we all care about. Look, when we needed to only solve for five events a year where I could just turn down my thermostat a little bit and hopefully reduce a kilowatt per household and bring us down to maybe a few hundred megawatts worth of load flex, that was one thing. Now when we're talking about doing that across so many hundreds of millions of devices to build an operational confidence which has cybersecurity, enterprise grade, data quality and control quality and do so in a way where a distribution operator can trust it as much as a bulk system operator can trust it, that's a different level of data standard that I can promise you we are not at and so how do we create that standard schema that we can all use? Because speaking to outcomes, I loved some of these stories, maybe I'll use a slightly more, how do I put it, slightly more stale but an important outcome which is we've got substations where we are operating our platform where we are going to push back the need to upgrade that substation by over a decade because of the amount of load flexibility we're able to deliver on that really hyper localized location and I think as we think about moving to, as you mentioned, hundreds of millions of devices, connected devices, we're no longer going to live in a paradigm where every customer is going to operate in the same way. We have to think about developing schedules where every individual is getting their own schedule for their needs, for their preferences, for their choices that serves the broader common function of supporting reliable and equitable transition so I think that's the way I kind of think about that. That's great and I mean when the Kyso pushes the flexler button, they're not just wanting to target General Motors cars, right? They want to put Ford and Toyota and everybody else to do that's how we get the kind of, so the state, so I'm always on the lookout for places where the state needs to kind of step in and get industry aligned on these kinds of things so just let's keep thinking about that and quickly if we could go down if you sort of have any response to the question. California can take a leadership role in this. The state has a lot of data that is, it can actually release, but it doesn't put it out in a machine readable form. It's in PDFs, it's in, you have to ask for it and so on. I think California can set the standard here. Yeah. I think also the spatial resolution of the data so I think it's common to have data at the county level, less common, like a census track level and oftentimes students are looking for data that is at that finer grain resolution. I think of course you have to balance privacy concerns with that and so I think there's a lot of interesting work to be done to think about how can you release finer grain data and maintain a high level of competence that you're preserving privacy. Yeah, we haven't talked too much about the data privacy questions but the ability to have some access to customer electric gas data is really essential if you think about decarbonizing buildings which is where all that energy is being consumed and data privacy concerns have primarily been the reason why it's so hard to work with and access that customer billing data but I think creative ways to aggregate and still provide granular data without compromising addresses and privacy concerns I think could just unleash a lot of innovation. We're gonna take this a completely different way. Great. The investment banks have figured out how to get interns to apply, hire them and get them into the pipeline. I have students that are equally as interested in saving the world as becoming rich and I would love them to go into saving the world side type things. Can we have a platform or somewhere you could call it the make the world a better place internship market or something like that where startups, government agencies and so on can go and get this sort of match up going so it's not this painful search process where the students then default to going into finance. All right, there's a Peace Corps for data. Yeah, I love that. I love it, I love it, I love it. Well, great, well thanks Super Insight Fly. We only have a few minutes left but I do wanna open up for audience questions. If you have any, there are two mics. I saw a gentleman right there and then that gentleman and then Claire. Hi, thanks everybody. I'm Alex Thorin with Linux Foundation Energy. And I'm curious to hear from the whole panel including Commissioner McAllister. Thank you for all of what you're doing at the Linux Foundation. That's great. You're welcome. What role do you all see for open source, open standards, open data to play in accelerating the energy transition? I'll answer that super quickly. So the Energy Commission in partnership with PG&E and some others funded over several years the creation of some algorithms that can take interval meter data, customer level interval meter data and really just rest a lot of knowledge out of it and understand where the seasonal loads are. Really sort of aggregate to small level, help local governments, that sort of thing but really get a lot of insight out of how energy is used and what loads might be flexible. Impact assessment of programs, expose that kind of thing. So there's just a lot of powerful tools now that didn't exist before and those now sit at the Linux Foundation because they have such social good aspects to them. They're so necessary that they needed to be out there for the public for anybody to use. And then a private actors can use it and add their bells and whistles and that's their value add but the basic public resource is there. So I think that's a perfect use of that platform. I don't know if anybody else has examples. I did wanna ask whether the commons is gonna be sort of or Khan Academy or something is gonna, you work with them to do a data science class or something. You're talking to them and a few others. Good, okay, great. But anyway, the sort of public open source aspect. Anybody have any insights? I think the social cost of carbon, the damage one ton of CO2 emitted does over eternity across all sectors globally used to live behind closed walls and then David Antoff and the EPA and a bunch of folks took it into complete open source out into the light. So now you can modify yourself and play around with it. So I think that is the wave of the future of scientific research is take the stuff out so you can use it, make it easily executable and if you've got a little bit of computational skill to edit it, but knowing where these numbers we use to make policy based on come from is absolutely key. So I'm two thumbs up on your question. Let me elaborate on that. That's an important, it's not just a nice to have, it's absolutely necessary. There's so many things in this space which are claimed by some people to be not quite true. And the only way you can shed transparency into it and get clarity is not by shouting louder but by making everything open and transparent and examinable by everybody. Here, here. Anybody else, Amber or yeah? No, okay, great. I mean, I think we do a lot of modeling and using open source tools. We work in Python because everybody can use it and I think there's a lot of benefits that come from letting other people fact check you or question this assumption or that calculation. Absolutely. So let's go on to the next question. Thanks so much. Seth Hiddle from the Post-Droid Foundation really appreciate the panel, loved it. I have a similar but I hope complimentary question which is there's real tendency, especially in the United States for vendors of technologies that go on the grid to have a proprietary interest in the data that their devices generate. How do we push back against that so that we can have open access to that data and sort of enable the innovation that we need to have applications and all the great things that we've talked about today occur? Great, I mean, that was kind of the origin. That issue was the origin of my question about sort of the standard, you mentioned the schema, right? And so I think the state has a role absolutely in that. Do we actually sort of go to the charge, the charger companies and say, okay, you have to at minimum comply with this protocol. OpenADR, we developed the Energy Commission, LBL did all that great work. So now we have a platform. How much do we require that? And say for GM, they have to do some minimum standard but then they can add whatever bells and whistles they want on top of that. So any more insight? Yeah, I mean, I can add a couple of thoughts. I think this is where I'm going to segment out the concept of data and controls. And I think the reason I will is it's odd for a panel that speaking so much about data, we haven't talked about cybersecurity. As all of these hundreds of millions of devices get onto the grid, we can knock out parts of the country's infrastructure very easily by hacking the right things. It is dangerous and we are entering a very dangerous year and I do not think that we are taking that seriously enough. And I do think that there's a little bit of a willy-nilly attitude to that where it's like, ah, it's fine, it's an API, I'll just talk to an API. No, that's a bunch of cars. Like, you gotta be really careful. I'm not by any means pushing back on open standards for data. I think that's very different. I think we should have an open approach to sharing data, sharing insights. But like, I'll give you an example. OCPP is an open charge protocol. It's not standard. There's no standardization. Every device manufacturer uses it totally differently. And so now you're stuck with trying to accelerate that transition but everyone's using it totally differently. So is it even open at the end of the day? Half the time when people use the word open, I actually think to myself, you're super closed, right? And then you layer on cybersecurity and it becomes a double threat, particularly when you wanna use the controls element. So it's an open question is what I would say rather than offering a pithy, yes, we will definitely do it tomorrow. Yeah, yeah. Any other observations we got? We do need to wrap up but any lovely nuggets you have would be welcome. Great. Well, I think we'll wrap up. I think there's one question. Oh, so let's see the organizers, do we have time for one more question? I'll just do a preemptive yes, okay. Claire Brew. Hi Claire. 358 area. And first of all, I worked for 28 years at CDC and I never thought I'd come to an energy conference and see CDC on the screen. But the question is sort of sharpening the two other questioners. Let's bring it to AMI data. When you're talking about data commons or access for third parties or CCAs, what is the appropriate way to deal with what I would call at best, data gatekeepers at worst, moats. And particularly I'm thinking about certain investor owned utilities where their contractors can get data with limited latency and CCAs can't about their own customers. So I'm interested whether you have solutions for that. Great. Well, that sounds like maybe that's a question for me more than anybody else. So actually we are working on that. There's no getting around the fact that PII, privacy data is attached to interval meter data at a customer, right? And so we have to be super, super careful with that. The Energy Commission now is collecting that data from all the utilities across the state. Certainly wherever there's AMI and where there's not. But we're on the same side of the firewall. We're not crossing the firewall out into the public to give that data away. We're protecting it very seriously. But their local governments, for example, really need that kind of data to do their climate planning, which they by law have to do, right? And so they're blind about their energy consumption and their jurisdictions more or less. And so we are working hard to solve that problem. And there's a kind of putting in place contractual arrangements that allow highly vetted firms to do analysis on behalf of say the local governments to rest the knowledge from that data but without actually showing the data outside of the firewall. So I think there's a, it's really promising. I've been working through, it's never been done apparently at state service in the state agency land. So we're sort of plowing some ground here, but we think that it is permissible. It's like contracting, the contracting law can handle it, but it's taking a little while, but we're definitely getting there. So great. Well, let's thank our panelists.