Dr. Teon Brooks 0:02 Oh, it auto-advanced. 0:04 OK, that was even before I was ready. 0:05 OK, cool. 0:05 I'm going to go back a slide. 0:07 Hi everyone. 0:07 My name is Tian Brooks. 0:10 This is a talk that I just gave at a symposium that I was leading last week. 0:15 I was trying to do some edits. 0:16 That did not happen. 0:17 So I'll be going through this again and updating it for this conference. 0:22 So as I said, my name is Tian Brooks. 0:25 I'm the president of the Gotham Data Clinic. 0:28 And we are a new nonprofit based in New York City. 0:31 We just received our 501(c)(3), so we are sprinting ahead. 0:37 So I'm just taking the show on the road and I'll tell you about some of the work that we're doing. 0:42 So that's one of many different hats that I wear. 0:45 So my background is in open source software development and I focus primarily on neuroscience research. 0:52 So I'm a neuroscientist by day and also My background's in linguistics and psychology, so I'm just in this weird cognitive science interdisciplinary space, which is neat. 1:05 So the job that pays my bills is the one that I work with the team at FAIR with Meta. 1:13 We have a research team called Brain in AI. 1:16 The Brain in AI team is focused on creating open source tools for neuroscience research. 1:23 And we just had a big announcement yesterday of some really cool work that we'll do, that we're doing. 1:29 I'll share some of that later. 1:31 But we released an open source package, an open source model. 1:36 And we're really working on trying to understand how the brain works. 1:40 So that's my daytime job. 1:43 The nonprofit is the thing that I'm very passionate about. 1:46 And we're trying to get this off the ground. 1:48 And they're all kind of tied together. 1:50 So I'll walk you through some of the work that we're doing with the nonprofit. 1:55 Okay. 1:55 So, 3 things that we focus on. 1:58 We do formative research. 2:00 We create tools and we do education. 2:02 And this talk is gonna be primarily about the education part and the speculative part of what we want to build from the research that we've done. 2:11 So, we recently did a landscape analysis of all the different ways that computing and data science is being taught at the high school level and in higher ed. 2:21 It's really fascinating some of the work that's happening in France right now. 2:27 The French government mandated that all of their high school students must learn to program, and that programming being done in Python 3. 2:35 But they gave them no budget and no guidance on what the regulations or needed to do so. 2:45 So imagine trying to scale a computing education program for hundreds of thousands of students, say half a million students, where the current state of the art was using JupyterHub. 3:00 And JupyterHub is really great, but trying to scale that for the number of students who just need to open a Jupyter notebook to type hello world, it's a bit of an oversell in terms of the capabilities versus what is necessary. 3:15 So, let me see. 3:18 Okay, I'll stay on this slide. 3:22 So, the idea that we have in the paper that we looked at was that there are some teachers in France, they came across this really cool project that's called Pyodide. 3:33 So, how many in the audience know about Pyodide? 3:36 Okay, we— I see a few hands. 3:38 The background about Pyodide is that it's the Python stack compiled to WebAssembly. 3:43 This is really neat because you can now run Python in the browser. 3:48 And if you can run Python in the browser, that means you can run it on a Chromebook, you can run it on an iPad, you can run it on any device that has a modern browser. 3:55 So this takes the computational load off of the servers and brings it directly into the browser. 4:01 Which is really necessary if you want to run a computing education program at scale. 4:08 So imagine you have half a million students that are online that you need access. 4:15 You can just, like, have the computers that they're going to be using. 4:18 That's the new computational environment. 4:20 And then you can create a very thin layer and have this be where you store files that are related to the work that they're doing. 4:29 This will be the Jupyter Notebook. 4:30 This can be assignments that they have. 4:32 So it's a really cool paradigm to start thinking about how you can actually scale education to actually meet the demands of the school system you're working with. 4:45 So we did a lot of really cool research. 4:47 There's a preprint. 4:49 I saw it on the slide earlier today, which is really cool. 4:52 Thanks, Ron. 4:53 So it's under review right now. 4:56 We just updated it. 4:58 Feel free to check it out. 4:59 I think there's in the slide— I'm just going to advance. 5:02 Okay, cool. 5:04 One project that we worked on for our nonprofit, so just fast forwarding just a little, is that we started a stewardship program for a project that we were involved with at NYU. 5:18 So at NYU, I helped lead the product development for this app called Brain Waves. 5:23 We were taking the paradigm of running experiments for neuroscience and psychology and bringing that to high school students in New York City, giving them an authentic research environment using really cool web technology. 5:40 So the app is built on Electron. 5:42 It uses LabJS to do experiment presentation. 5:47 It also can use JS Psych, which is another experiment presentation library. 5:51 Then we actually implemented using Pyodide to do all of the data analysis. 5:58 So now you have an app that can run on a computer and you can show a student from soup to nuts, from the beginning to the end of like how you can develop and design an experiment, run the experiment. 6:12 The cool thing about this I forgot to mention is that we connected it with an EEG. 6:18 So you can do EEG over Bluetooth, so we get to show students their brain activity for the very first time streaming directly into the app, and then they're able to do the data analysis using the same tools that I helped build in Python for the data analysis. 6:34 MNE Python is a really cool library, you should check it out. 6:37 So we developed this, it was under a grant for 5 years, we wanted to be a steward when the grant finished. 6:43 So that is what born the nonprofit. 6:47 So we really got into neuroscience education, and we realized that this would be a great opportunity to just get more students into the neuroscience field. 6:59 Then 2020 happened, and then the whole world shut down. 7:03 So we decided to look at other opportunities, and we decided to extend our neuroscience offering computing and data science because you can do that on the web. 7:16 Sorry. 7:18 You can do that on the web and you can scale to the students you're working with. 7:23 So I'm going to fast forward. 7:25 So that's the motivation behind the nonprofit. 7:28 I touched on this slide. 7:31 The really cool project that you can check out in France is called the Capitao project. 7:35 I'll also just try to share some links later today. 7:39 It's in the bottom right. 7:40 You can see that the CopyTile project is involved with the Pyodide and also the Jupyter ecosystem. 7:46 So it's a really great opportunity to bring all of these different tools into an ecosystem that you can start scaling computing and data science education. 7:56 The French model is really, really cool. 7:58 And I'll wrap this up because I don't even know what's in the rest of the slide deck at this moment. 8:03 But OK, the French model is really cool because it allows for you to create computational notebooks, store it very much in a low-resource way, and you can serve this to hundreds of thousands of students. 8:20 So the idea that I have for the nonprofit and what we want to work on is taking this whole model that we have seen from the Friends School and bringing this to the app protocol. 8:31 So the app protocol will be very interesting in terms of like work that's being done with Tangled. 8:36 So Tangled, if you haven't heard, is an open source alternative to GitHub. 8:41 So you can create these different environments where you can like store information and you can clone that data or you can fork it. 8:49 So the nice thing about this model is that you can have attribution for the work that's being created and you can have educators create the material on the app protocol. 9:03 You can use it as an identity layer, you can have it as an attribution layer, and that's some of the work that we want to do with computing and data science education. 9:12 Create notebooks that can be stored there, teachers can create that material, they can share it out more broadly. 9:18 So yeah, I will stop there. 9:20 There's the paper you can check out. 9:22 There's some tools that we're going to be building. 9:24 I would like some help because I've never built anything on the app protocol, but I built other stuff, so I would love to, like, combine forces with anyone here, and I'll just leave you with those thoughts. 9:34 Thank you. 9:35 Thanks. 9:40 Thank you, Tion. 9:41 Are there any questions for Tion? 9:49 No questions. 9:50 Everything clear. 9:51 Cool. 9:52 Yeah. 9:52 That's, that's fine. 9:53 You can always find me later.