 I don't want to lose sight of this question. What is big data actually for? So as we're thinking about harnessing the data revolution, we need to think about why we're actually harnessing it and what we're actually harnessing it for. My name's Jennifer Balch. I'm the director of Earth Lab here at CU Boulder, and I'm also the university director of the North Central Climate Adaptation Science Center. Big data is for humanity, right? We have essentially justified sending up dozens of space satellites to improve our living conditions, to improve people's lives. And not just some people's lives, but all people's lives. I would argue that being able to access data and use data is a basic human right, that every kid needs to learn a second language, and that language needs to be a programming language. The next great exchanges of ideas are gonna be happening in code. But we sit at a crisis moment. The data divide has the potential to get bigger and bigger and bigger. Given the amount of data that's being generated every single day. But there's also an opportunity in this. It could be a leveling. Open science, open data, open access could potentially harness the diversity revolution. And in fact, in my own situation, I was able to have babies and data too. It enabled me as a woman in science to actually do an incredible amount of fire science based on the fact that there were satellites capturing information about fire and doing the hard work, or part of the hard work for me. We have a diversity of data. This is the big data problem. This is just the quiver of satellites that NASA has pointed at our Earth. But we also have a variety problem. And underneath that, particularly for the Earth and environmental sciences, we've got a heterogeneity of data types. So one example from Earth Hub, we're matching satellite data on fire and floods and hazards with 200 million housing records from Zillow. Or social media data. Millions of tweets around hazard events. But with heterogeneity, we need to have a diversity of people actually helping us understand where those data are coming from and what they're actually representing. And we need somebody like Adon to be on the mission development teams at NASA, right? All of these satellites have bias in their spatial and temporal distributions and the spectral reflectances and signatures that they're actually capturing. And we need to recognize that, anticipate that, and accommodate that. We also have a diversity of complex challenges. There's no way one single person is gonna come up with the solutions. So there's a demand right now for large team science, not only within academia, but cross sector, bringing in multiple disciplines, perspectives, ideas, career stages to the table. And if we can't cultivate the conversation that needs to happen, how are we gonna possibly solve our complex environmental challenges? So fundamentally, we need this conference. We need a diversity of people working on the solutions. So we need this conversation to be happening. And I would argue that includes, which is one of NSF's 10 big ideas is actually central to all of them. We can't actually navigate the new Arctic. We can't harness the new data revolution. We can't do convergence research. We can't think about what the future workforce looks like without incorporating diversity as central to that. And the reason why I say this is because diversity is key to innovation. We ultimately need a diversity of perspectives because it's at those friction points, those healthy friction points where new ideas are actually generated. What are the solutions to this problem, right? So how do we increase diversity and inclusion in data science right from the start? I think there are three that we should think about. One is that we need to shift the culture of how we do science. We need to do co-production. We need to think about end users, stakeholders. Who are the people that this is benefiting? Are they engaged? Are they participating? Are they contributing? Along the spectrum of data discovery and decisions. The second solution is open science. We can't just build it and let people come. I get that, but we do need to also change the culture of how we're delivering and sharing what we do. We need to make our code, our data, our workflows and entire process and tool development open and reproducible. Our third solution is that we need open education. Have to be making our trainings available to as many people as possible. 700 PIs in bio said the biggest gap they see is in training. At Earthly we've developed one of the first national programs in Earth analytics, which is a blend of data science and Earth systems knowledge. Check us out at earthdata science.org. We have three full courses available, 250 training modules. We are now getting 41,000 global users a month. This site is about six months old. Just indicating the rapid need and demand for these skills. Thank you.