Transcript#

This transcript was generated automatically and may contain errors.

Hey there, welcome to the Paws at Data Science Hangout. I'm Libby Herron, and this is a recording of our weekly community call that happens every Thursday at 12 p.m. U.S. Eastern Time. If you are not joining us live, you miss out on the amazing chat that's going on. So find the link in the description where you can add our call to your calendar and come hang out with the most supportive, friendly, and funny data community you'll ever experience. Can't wait to see you there.

And I'm so excited to introduce our featured leader today, who I think is building a community of learners that's really amazing. We have here Michael Czahor, who is the President in Athletics. I'm going to let him introduce himself. Michael, if you could tell us a little bit about yourself, what you do, and something you like to do for fun. I'll try to remind you of that part. It's hard to remember a two-part question.

Yeah, thanks. Thanks, Libby. Hopefully you can hear me okay. Yeah, so my name is Michael Czahor. As Libby said, I run a startup called Athletics. We recently partnered with Posit to include some of their tools on our site. My background is largely in statistics and reliability engineering. I went to school at Iowa State University, which I'm actually like a mile from campus right now. I'm visiting my in-laws, so I'm pretty close to where I went to school. I have spent most of my career in sports analytics. I have worked for several different professional MLB teams. I started as an intern in the NBA. And then I currently also work for a tech company working on high-speed camera technology for performance science.

But beyond that, I have spent a lot of my focus on teaching. And so I started this company back in 2023, as kind of like a second iteration of something I had been working on. It started off as just teaching data science using sports concepts, but it grew into more of kind of a technology platform to have people be able to streamline data science into a learning management system, which I'll talk a little bit about today. And so we've integrated a lot of Posit tools directly to our learners in real time. And a big reason we wanted to do that is because we feel that the e-learning space is moving really quickly with all the AI things that are going on right now. And we wanted to be able to spin up content faster and faster, and that required a solution similar to the one that I'm going to share with you.

And we currently have about 25 different instructors involved, focusing on a bunch of different topic areas. And we've had over 1,000 students take our material in the last three years. And we are looking forward to talking a little bit more about what it is that we do. As far as what I like to do for fun, I definitely like to go watch a lot of live music. And so myself and my wife, we live really close to Long Beach Island in New Jersey, where there's a lot of different live music going on all the time. And so we like to do that. I would definitely like to travel. I have two kids, one and three years old. And so we're definitely in the busy stage of life right now. And so a lot of our days are spent with both of our daughters. And so definitely enjoying that as well.

Amazing. Okay. What about something that you like to do for fun, completely outside of data science?

Yeah. So outside of data science, definitely like both attending sports games and coaching youth sports. So I spent a lot of time during graduate school coaching girls soccer, which I coached for five years during school. And then prior to grad school, I coached youth baseball for about five years. And so something that I'll probably get back into, haven't done it in a couple of years, but really enjoy doing things on the coaching side. And so yeah. And then like I mentioned before, traveling, whenever we go somewhere, we try to do like a combination of both like the music stuff, like I mentioned, but also whatever like historical stuff is in the area, we'll try to go around and do a little bit of that. And so definitely liking to see new places. I've definitely done a lot of travel just over the course of the last few years, just from working for, for teams. So I've got to see a lot of different cool cities. And so I thought the travel would slow down a little with life, but it really hasn't. It really hasn't. So I've, this, this, this particular last two weeks has been kind of crazy. I was just in, I was just in LA for work last week. I just completed a 20 hour drive as of like yesterday, yesterday out to my in-laws. And then I'm actually flying to Boston tomorrow for JSM. So if anyone's going to be at joint statistical meetings in Boston, we will have a, we will have a booth there. I should probably know it for this call, but I think we're booth like two 11, but come in, come and say hi. I'll be there. Yeah. I'll be there from like tomorrow until Tuesday, I think.

Sweet. Well, I know Donald will be there. Donald just joined us. You can see him walking on his treadmill. And I know a few other people will be there. I want to say Elizabeth Eisenhower from our crew will be there from the do sides, hang out kind of attendee crew. And she'll be joining us on the hangout in a couple of weeks too, as well. Okay. If you are going to JSM, everybody go find Michael and say, hello. I already have questions that I want to ask you. I feel like we should hop into Slido because I selfishly want to ask you about coaching and how your experiences coaching have informed your teaching. Cause I know I'm sure that they have, I love coaching too.

Yeah. I can maybe give like a quick note on that. I would say, I would say like on the coaching side, I mean, you similar to, I mean, similar to teaching you're dealing with either students or players in a variety of different backgrounds. And so you really kind of have to be receptive to kind of where people are coming from, both from like a skill point, both from like a skill point of view, but also just like a life point of view. And so I've coached, I've coached players from a variety of different backgrounds, variety of different talent levels. And I would say just continuously getting feedback from the player on like how things are going, trying to individualize the coaching points to them as much as possible. It's always good at, you know, if you're coaching a soccer practice, baseball practice, whatever it might be to have kind of more you know, universal things that you're showing. But at the end of the day you definitely have to tailor it towards the student or player. And so I have taken a lot of time over the years to try to get things right with how I plan sessions. Whether that be from coaching in sports or from teaching in the classroom. And it definitely depends on the size of the group that you have. It's obviously easier to hyper focus on a group, whether that be players or students, if you have three to five people versus if you have 80 to 100 people in a room. And so I definitely approach those things differently. But I think it just kind of comes with, you know, being as conversational as possible with people, both during and after the sessions. And so I've taken a lot of time with students, you know, outside of the classroom, even when I was TAing in school, just to try to learn what it is that they like and try to give examples that help them learn better.

Amazing. I think that's what you're doing at athletics. Well, I would love to start with a anonymous question that came into the Slido, because I think it's very interesting as someone who is teaching and trying to stay at the edge of the, like the front of the curve of what people actually need to know to go into industry. This question says, which tech trends are actually worth the hype? And which are just shiny objects?

Yeah, yeah, I, I've, over the years, whether it's for work, or just doing random things on like LinkedIn, or blogs, I've, I've always enjoyed making my own applications, in terms of like, whether it be like a shiny app, or react app, something like that, to kind of like, you know, hopefully get people engaged. And those things are so easy to make right now with all the different AI tooling that's going on, it's very easy to go on and kind of just type into a prompt and say, go ahead and make me an app. I think the I think the tooling that is worth knowing these days are tools that are getting you to the point of being able to produce something that's production level faster, but still understanding that things need to, you know, go into production. And so I see a lot of people that will spin things up, and they'll host it, and it might just be like an HTML, you know, an HTML based application, and then getting that, getting that to be able to be used day to day is a lot more difficult than just kind of putting up that demo.

And so some of that tooling that I've used for that, I use a lot of cloud code, cursor, different coding assistance to kind of help me speed up, I don't want to say the boring parts, but the parts that are kind of used on every single project, whether that be exploratory data analysis, or being able to prototype a visualization faster. And so some of those some of the tools, at least on the LLM AI side, whatever you want to phrase it, I'm using much more for prototyping than I am implementation, at least at the base level, I think once you get a little further along, there's there's other things. As far as the different subscriptions, though, that I see floating around different places. You know, everyone has a SAS, everyone has a SAS product now. And it's, it's really easy to kind of, you know, buy into different buzz pages, whether it's like, you know, here's a here's a teaching tool that I'm using that someone could have designed in, you know, a day, it's hard, it's hard to know. And I think I think the way that you evaluate that is just making sure that reputable sources are using them, that the use case isn't something that you can solve yourself very quickly. And so I've definitely just tried staying with some of the more mainstream ones that are being used, you know, more globally, but I'm all I'm continuously I'm continuously looking for, you know, refinements and, you know, how to keep my toolkit as strong as possible.

For what we're doing at athletics in terms of like, how we're streamlining people to get onto the system with with posits with posit stuff. You know, we, we were very, we were very excited to be able to kind of bring people, you know, a cloud based version of, you know, posits workbench on our site. And, you know, we already we already, you know, had a relationship with folks who have been working with posit. And so we felt good about that. But as far as like, the question goes in terms of how do you how do you know what's just a kind of a shiny object versus, you know, potentially something that's worth doing, I definitely think it's talking, talking to the people who are using it at the highest levels. And so any instructors or folks that we're bringing in, we're continuously asking them what they're using in their job. And we're trying to we're trying to understand because you certainly can't know everything. And that's why we're trying to have as kind of a much of a reach as possible, not just in terms of our own information, but what others are using as well. And so we're evaluating this kind of stuff every day.

I definitely think it's talking, talking to the people who are using it at the highest levels. And so any instructors or folks that we're bringing in, we're continuously asking them what they're using in their job.

Yeah, well, I mean, I think it helps that you are bringing in people who work in various places in industry to be teachers or mentors, and leading classes. I don't think that she's here today. But Verla, who a lot of people know from all of her shiny teaching and everything that she shares about shiny online works with you a little bit. And I know that you have a whole host of folks who do this. I think that's amazing. So we, we had a couple of questions come in the Slido. For anybody who is not watching the chat, we have lots of jobs as well, like things that are hiring. So take a look. Rachel just posted a few. I know that Javier, who hasn't joined us in a while, is hiring as well. So check out that and really share those with people to help them reach opportunities.

Audience Q&A: sports and data

Okay, we had a question from Adam. Adam T. Adam Taggart, I believe. If you would like to see if we can get him unmuted and ask that one live, that'd be great.

Oh, Highlight it. There he is. Hey, Adam.

Oh, everyone. Hi, Michael. Thank you so much for being here. And also what a massive relief it is to hear a fellow coach talking about personalizing coaching to the athlete, to the student. And also, you know, I think people who do that for their local community should always receive thanks because it is a lot of time and effort for not always a great reward. So I think that's a wonderful thing. Thank you for that. My question is about the usefulness of the data for different sports and different disciplines and skills and stuff. And obviously baseball is the one a lot of people know from sabernomics and moneyball and things like that, where there's all of these different metrics, time on first base, time on second and so on, which allow that to generate useful data towards increasing performance of the sport. And also we were talking to Jamie not too long ago about hockey and the fan analytics and how that is useful data towards people getting interested in the sport. My question is, what, in your opinion, are the easiest sports to generate useful data for? And what are the ones which it's damn near impossible, seriously difficult to generate useful data for?

Yeah, thanks for the question and for the nice comments about coaching. I would say in terms of the, I guess I'll take it two ways in terms of generating data. I would say there are sports, I'll start with like what sports have are easier to do projects on? I have found, I have found that sports that are more discreet are generally easier in terms of the way the data comes in to be able to kind of make some sort of meaningful contribution.

So I was lucky enough to work in baseball for several years and baseball is a very discreet sport in the sense that every single pitch is kind of its own play. You're able to kind of make evaluation at the play level where as other sports, perhaps something like soccer is a bit more continuous, the game is kind of more free-flowing.

And I've definitely found that being able to make insight on the discreet sports for a variety of reasons is probably much more intriguing, at least to me, as far as the actual data collection for some of these sports. A lot of sports are getting much more into tracking technologies, which is happening quite a bit. You just mentioned, I missed the name and I should know it because I've seen her post, but the woman from the NHL, I'm sure she probably spoke about some stuff perhaps with tracking in the NHL that is happening now. And of course, in baseball, back around 2020, some of the Hawkeye information began to come around, which is, it's essentially like an 18-point tracking system where you have essentially like markerless information on players at a very high frequency. And the question becomes, how do you actually make sense of this data? What's actually useful information? And what are these teams actually doing with this? And what's worth knowing?

And once this tracking stuff started coming around, a lot of the teams, their minds immediately went to, and not just for baseball, I think this was kind of more globally across sports is how do we use this for things like injury prevention and player health and trying to better understand different movement patterns and things like that. And I think depending on who you talk to and how far along teams are in a given sport, different folks will have different answers for how they're using it. Some might just be using it just for exploratory analysis, and they might just be making really cool charts to try to understand what players' different key performance indicators are in a given game. Others might be making joint models for longitudinal and time-to-event data and trying to understand what different behaviors might lead up to some sort of injury.

The sports that are benefiting more right now from different, like tracking options and computer vision and things like that, or at least the ones that are interesting to me are the ones that are not as major of sports. And so things maybe like gymnastics might be an example where you're able to use a lot of different vision technologies to try to understand player movement in the game, where they might not necessarily have a lot of the camera tracking that some of these other sports have. So I find that to kind of be interesting. And I think the more unique movements that players are making in those sports, the harder it is to understand. And so I definitely think if you talk to folks that are more on the biomechanics side and player development side for specific sports, they would tell you that for their sport, they're looking at very specific things, whereas, you know, baseball, they might be looking at overhand throwing motions for players, whereas in something like rowing, they might be looking at something different for players. And so it definitely varies, but I would say that the data collection has never probably been richer for sports, given a lot of the technology that exists right now.

I would say that the data collection has never probably been richer for sports, given a lot of the technology that exists right now.

And, you know, kind of turning it back to some of the stuff that we're doing, at least with the teaching at AthlyticZ, I would say we are really big on showing folks a lot of these data in our different modules. And so we're getting students to work with some of this high-frequency time series data. When we don't have injury type data, you know, available, we might be showing students how to kind of simulate event-based data and then try to connect the two pieces. Because I think a lot of that stuff is super interesting, and it actually, you know, helps students learn different kind of methods of how to evaluate it, whether it be machine learning methods, Bayesian methods. And so, you know, it's going a little off your point, but I think, you know, it's extremely interesting to try to learn through those types of data, because they're really rich in information, and it has a lot of cross-connection for different areas of data science and statistics.

Questions from the audience

I think this is bang on for a follow-on question. It was anonymous on Slido, and I would love to know, too. This anonymous question says, hello, I actually have a teaching- centric background and got into data analytics after. How are your learning materials organized and curated? How do you manage your resources, and are the sessions in-person or virtual, and do you ever hire for data science instructors? So this is a sneaky four-part question. Let's start with the first one. How are your learning materials organized or curated? And I think that this probably means, like, within the context of the courses that you offer and things like that. Also, I'm really excited to hear that you have this, like, plethora of data types that let you do different types of analyses. I don't know a lot about sports data, so that's fantastic to hear.

Yeah. Yeah, thanks, Adam. I just saw your comment there. Thanks for the question. I appreciate it. Yeah, in terms of how things are organized, I would say one thing that definitely differentiates us, and I would, you know, kind of pitch ourselves to say, please stop by at JSM to see this in person. We, along the way, we built our own learning management system, which is not something that everyone that's doing content creation is doing, and for those that don't know what that means, if you are currently in academia, you might be using something like Blackboard or Canvas, and if you don't, and if you're not in academia, you might see things online where companies let you upload your own content, and there's nothing wrong with those companies. They're extremely efficient. They let you post a video and then type HTML or just have widget-based things to go into it. We built our own, and the reason we built our own is because we wanted to be able to streamline true data science environments to learners. We are a data science-centric learning management system, and so we are organizing all of our content directly through that LMS.

We are definitely leveraging a lot of AI for our notes at this point, but the way that our actual instruction goes, instructors are encouraged to take their own teaching styles, create their content. We do have some standard operating procedures for how they should submit things like code to us, but in general, those instructors within their own disciplines are encouraged to organize their slides, their code, record their videos, and then they send them to our team, and then we eventually put them onto our site, and that's more for the courses. As far as how do we actually outline a syllabus or a curriculum or something like that, we're working directly with the instructor. We are researching literally any university or public offering on the e-learning side that you can think of and making sure that we're meeting the standard of what students are learning in college. We try to be extremely academic on our site while still being an e-learning site, and so we are researching pretty much everything that exists at the time that the course is created.

This kind of brings me to my next point of, you know, okay, as you create a course on say July 30th, what happens a month, a year from now? Do these courses go stale, and how do you manage that on the site? We just recently launched what we call AthlyticZ Masterclasses or AthlyticZ Live, and this is kind of what led into the partnership with Posit, was we needed to be able to spin up virtual environments within like five seconds. It takes a long time to do what we were doing for courses, and so we now have the ability to take a zip file from an instructor where they are either teaching in R or Python. We drop it onto our content management system, and then a virtual environment will deploy, and then it lands directly on the page where the instructor is teaching, and so it lets us deploy these really quickly.

As far as how we track things like student progress on there, we have a bunch of different analytics tools on the site to track completion. If there's quizzes, student progress, and things like that, but I think really the most important thing that we're doing for organizing all the content is we are making sure that the code runs before the session, and that's getting a little outside this question, but having taught both to students while TAing or doing my own stuff on the consulting side, one thing that always comes up is you might get a contact from a student, hey, my environment doesn't work, hey, I'm having install issues, hey, I'm having version issues. There is no session that runs on our platform that doesn't have that solved before the session starts, and it's something that we plan on bringing to conferences and to workshops and universities between about a month ago and 2027. We want to be able to have people show up to those things, use professional environments, and so that's definitely something that we look forward to doing, and if you're at JSM again, we'll have demos going on this throughout the entire event.

Day-to-day life balancing industry and AthlyticZ

Amazing. Okay, I would love to ask the Natasha question, which is wonderful on Slido all the time. Thank you, Natasha. It's what does an average day or week look like for you? I know that you do athletics, but you also work in industry, so maybe not an average day or week, but even maybe an average month.

Yeah, so my day-to-day right now is generally, it's predominantly my day job, which is I work for a company out in California doing the high-speed camera stuff, and that takes up probably between, depending on the day, probably between seven to ten hours in a day. Right now, I have other folks who've been helping me run AthlyticZ, anywhere from the instructor, instructor onboarding, making sure the sessions run well live, and so I would say AthlyticZ has a lot of autonomy compared to what it used to, and so if you asked me this question maybe two years ago, it probably would have been a little bit of a different answer in terms of balancing the two things. And definitely on the weekends, I'm definitely trying to just kind of check in on how sessions have gone, especially now that we're running live events, and so I might do a little bit of video editing on weekends, might be doing a little bit of just

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout Michael Czahor, Data Science Hangout

stay as curious as possible because right now there's never been a time to where it's been easier to be innovative with a lot of the AI stuff and I would just say from an academic side and just a learning side please do not ignore the fundamentals of anything that you're doing on mathematics or statistics I try to tell any students that I'm talking to to not skip over the the basics it helps you ask better questions it doesn't mean that you're gonna be I'm not encouraging people to do proofs at work and derive things from scratch but I think if you can ask a better question because you have a better understanding of something I think it'll make you a much stronger data scientist and statistician well said thank you so much Michael well I heartily agree talk to people talk to people at the data science hangout talk to people in the discord like Rachel said go to posit conf and meet people and I highly encourage you to talk about things that are not data related I was talking with no tobacco the other day about hand sewing and embroidery and how meditative that is we are whole people who do more than just data science be a whole person and recognize that in others go talk about fun stuff go ask Connor about birds I love talking with Connor about birds yeah hey Connor okay everybody this was so much fun you can save the chat if you would like to save all the resources or the LinkedIn links that people have shared there's three dots in the top right corner and we hope that you will join us at the data science lab which is on Tuesdays and again here if this was your first time we hope that you will come back and join us again we always have fun conversations going on and I believe next time we are going to be joined by Corey Campbell from NIH are so if you are in the healthcare space or curious about that come meet Corey she's gonna be super super fun all right everybody I hope you have a wonderful afternoon evening morning wherever you are in the world and we will see you again next week

I think if you can ask a better question because you have a better understanding of something I think it'll make you a much stronger data scientist and statistician