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. I am super excited to introduce you to Rylan Murray. Rylan is the Manager of Data Analytics and Development at Nebraska Methodist Health System. And Rylan, I would love it if you could introduce yourself. Tell us a little bit about what you do at Nebraska Methodist Health System, but then also something you like to do for fun.
Yeah, sure thing. So I currently manage data science, data engineering for the health system, but I spent the beginning of my career as an individual contributor data scientist. And I've been using RStudio for the entirety of my career, actually, since college. So the last 10-ish years, I've been a pretty avid user, and it's stayed with me into my career and into our organization now.
And for fun, the last three months, I have a three-month-old right now. So the last three months, there hasn't been a lot of that. But prior, it was, last year, I ran a half marathon. So running, working out, and then I'm kind of a serial hobbyist. So recently, it's been some home labbing, many PCs in my house, self-hosting some things that are useful. And I also have been in a five- or six-year Dungeons & Dragons campaign. So a little bit of everything. Serial hobbyist of sorts. Has your D&D game paused since you had your keto as well?
I've been calling in more so than attending in person. So I'll usually hop in once my oldest is asleep around eight, then I usually call in or hop in our Discord and join that way. But the rest of the group's still in person. So it's a little different, but a little easier that way.
It's so tough. Anybody who does a game night, whether that's D&D or anything else, knows that, gosh, it's really hard as an adult to make a weekly thing happen. So kudos to you. I know that myself and several other people from the Hangout are all D&D people. So you have lots of friends here. Rowan, Jared, everybody's chiming in in the chat here.
Building a code-first data team from scratch
Okay. So I would love to chat with you right off the bat about just orienting ourselves around your tools. Because you mentioned RStudio. I learned RStudio in school, and then I brought that love with me. I would love to know in the chat if anybody else has done that. Did you bring your love of your programming language and or your IDE to your job? Because Ryland, I feel like your story is unique. You got hired as sort of the first data hire in your organization, which meant you got to have some say over what tools were used and the platforms were used. Could you talk a little bit about your journey from the very beginning there?
Yeah. So I was a math major, and math and computer science to start with. I was going to be an actuary. And then we introduced a data science concentration during my undergrad. And that was exclusively built around the R programming language and using RStudio. And so my junior and senior year of undergrad, it was all learning how we can work with data, how you can programmatically automate data processes, visualize that data, and kind of just making sense of it using R and RStudio.
So then, yeah, the health system here locally was hiring for a data science intern. They really hadn't ventured in to that space, but data science was kind of a hot topic at the time and wanted to see what that would look like. They brought me on, and they basically said, we have no budget for tools. What can you do? And I just knew RStudio and R at the time could do a lot of what they wanted. They were in Excel mostly. And so there was a lot of low-hanging fruit early on to just set up R scripts, markdown files. We were really heavily into the flex dashboarding world at the time because those could just kind of be ran and automated and didn't need to be hosted. And that's kind of grown now into a team built around both R and Python now, but leveraging Shiny applications as almost pretty much our enterprise BI tooling stack at this time.
That's pretty amazing because I feel like a lot of people come into an organization that's already heavily invested in point-and-click BI tools, and then they sort of have a struggle to battle against that and bring code first in. So it sounds like you got to do that from the beginning. Is that part of how you got to build your team as well? Did you hire people who were already proficient in R or Python and already really in a code-first mindset, or did you train people up? I know you started as an intern. At what point did you transition into IC manager building a team?
Yeah, so I spent about three years in the IC role. Obviously, post-grad, I went into a full-time data science position, not the internship anymore. But I spent about six months. I had another colleague at the time. We had another intern, and then we both went full-time. But shortly after they left, it was just me for a while again. But the more steam we got with those, like I said, the low-hanging fruit, showing people that they could get a lot more out of their data than they could with the manual Excel work, we kind of were able to start bringing on new hires.
And luckily, one of our local colleges here was that undergrad program. So we kind of started off feeding off of that. We brought some two of our current data scientists came from Nebraska, Omaha, the college that I went to as well. And that program really focuses on that skill set. So they already kind of had that coming in. And then our third data science hire, we did have to train up a little bit. They came from like a business analytics program, but the critical thinking problem solving was there. So upscaling and decoding, I feel like was pretty easy, especially, I mean, anymore, decoding is almost the easy part with some of the AI solutions out there. It's more the having those problem solving critical thinking skills. That's really what I've always looked for in my team. And so upscaling a little bit, maybe on the hard stuff, but the soft skills were kind of always there.
It's more the having those problem solving critical thinking skills. That's really what I've always looked for in my team.
Sweet. That's great info for anybody out there who is looking for a job right now and wondering if soft skills still matter. Yes, they absolutely do. Problem solving is still a skill, especially one that I think we all might be dulling a little bit by using AI. So the more you can prove that you still have that in your brain, maybe the better off.
A typical day and the data the team works with
Well, I wanted to ask Natasha's question from our Slido, which will help us get oriented as well with what you do and how you work. Natasha asks, what does a typical day or week look like for you? What areas do you split your time between, like programming or meetings, et cetera? Yeah, so I'd say the last couple of years, it's definitely been sliding more into meetings. I would say I was closer to like 50-50 for a while, meetings and just hard coding work. But anymore, it's become a lot more, you know, I work with a lot of clinical leaders throughout our, we have four hospitals. So I meet with a lot of clinical leaders in those different areas and help them collaborate to make sure that we're, you know, if one hospital is doing an initiative, the other hospital should probably be involved, you know, centralizing the data for those different domains. That and then running some of our various data governance committees. We have an AI committee now and that's fallen into my lap as well. So a lot of committee work and stakeholder meetings anymore, and then trying to get better at delegating the actual project work to my team. That's, they're very highly capable. It's just that kind of like a separation thing for me a little bit, but more 70-30 now meetings to work.
70-30 now meetings to work. Okay. Curious if anybody else who went from IC, which is individual contributor over to the leadership side, what your kind of breakdown is. I think it would help to ask you as well, what type of data that you see, because you mentioned hospital administrators and, or maybe medical staff clinicians. I was wondering what kind of solutions your team is working on or building and what type of data that you see without needing to tell us like AI or anything. Yeah. Yeah. I mean, it really varies from, you know, the most discreet, like, you know, oxygen levels on patients that are ventilated over time, trending dashboards, things like that. Very discreet clinical data all the way to, you know, daily revenue reports, financial analysis. We work really closely with all of the different, pretty much any domain you could think of in a health system or just any organization from operational finance to clinical, you know, my team assists with. So there's not, that's, that's why I say keeps it the most intriguing is that it's not just the same old, you know, same old domain, you know, every week and every month brings kind of a new, something we're not used to looking at a new topic.
Getting credit for your work
All right. Well, we better hop to Slido questions cause they're coming in fast and furious. Oh, they're moving around so much. Okay. You can upvote your questions in Slido, by the way, there's little like thumbs up things that you can. So if you have a question, you want to read the ones in Slido first and see if anybody else has asked yours. There's also a little reply button. I hope, hopefully I've put on replies. I think I have that will let you add onto a question if you see when you like. But here it is. We have a question that was anonymous and it says, do you have any advice for making sure your work gets recognized? I found that once an analysis gets circulated or incorporated into a report, people no longer know it was me that produced the analysis.
Have you ever handled that in your career? Yeah. So we actually, so most of our stuff now, our analyses are shiny applications or Markdown dash. It doesn't really make that big of a difference how it's deployed, I guess. But in our sidebar, we always put you know, the little TLDR of what the content is. And then, you know, please contact my whole team actually just puts, please contact Ryland Murray with questions on this for better or for worse. We kind of just put like a footnote or a note that says like, and it's not really saying like, this is mine, but it says kind of who to contact. And that kind of gives enough ownership to, so that people know where it came from, both to route questions and concerns, but also to make sure that nobody's trying to necessarily take that work and spin it as their own. That obviously doesn't stop someone from screenshotting it or, you know, taking it into their own slides. But, you know, that's going to be a harder problem to solve because that seems like a little bit of an organizational, you know, ownership issue than anything else. But definitely, you know, putting that in just like in the HTML code or anything like that on your sidebar, a footnote, just something saying where it came from, that's been the most successful thing for us. Like I said, both throughout, so people know who to contact if they have an issue or, and also to make sure people know, you know, what team's working on it.
I like that. Cause I feel like some people might be a little shy about like, I don't know, putting out, Libby made this, this is my work, don't steal it, whatever. But like really making it as the point of like the, you know, who to contact if you have questions about this, I feel like is a great way to do that. And I have seen that a lot as well. Like if you have questions, contact this person or in this team at least. Yeah. And also early, just like one other thing on that early on my, my manager, when I was an intern, you know, he offered to, you know, if I wasn't comfortable going to the meetings to demonstrate or show it, but I think me, you know, taking the initiative to say like, no, I would like to make sure I present that getting your face on the work and actually speaking to it. That did a lot for my career as a whole, because people started attributing the work to me specifically versus, you know, my manager going to present it. So that did help a lot, especially early on. Origin favors the bold, right? Also like invite people to ask questions, which is what sometimes with the dashboard might be fraught because everyone will want something different from your dashboard. But then you get to talk to people and they get to ask you questions and know more about you. And that is always a win.
Getting buy-in for open source tools
Okay. I wanted to extend that question just slightly and ask how you got buy-in because you started as this first data person. You had another person you got hired with. It sounds like you two were on the same team going the same direction, which is awesome. And then you started with free tools, which is how a lot of people start, right? Like all of our tools are open source. Anybody can use all of our tools that we make at Posit Open Source. How did you get buy-in eventually to be like, okay, you've seen what we can do. Now let us have like the slightly more expensive version of this so that we can do it more safely and like scale it. Yeah. Yeah. So I mean, I think demoing, especially to my leadership, like the advantage of like a shiny application compared to a static, you know, flex dashboard, that's an HTML file was the most like that got us the biggest, I guess, investment from going from just our studio, no connect, no workbench to, you know, well, if we use connect, then we can publish shiny applications and then they can be, you know, real time based on the queries underneath. And you can filter and see like dynamic changes, like just kind of demonstrating that. And then also it not being a super expensive tool, even after all that investment, that kind of helped me like justify the ROI a little bit, but really just being able to say, well, here's the value add, here's the cost. I think we can easily justify that. That was probably, you know, the best way I went about that and got a lot of buy-in at the time, especially right. It was right around COVID when we did that too. And we had some COVID dashboards that really added a lot of value. And so that kind of helped expose, be like, well, this is what we're doing now,
because that's the tools we had. They were scheduled hourly markdown files and then saying like, well, if we had, you know, the connect and we could publish shiny applications, it would be even more dynamic and more real time. And you could dive into the data even further.
So, so it sounds like the answer to both the, like, how do you make sure people know that it's you working and how do you get buy-in for stuff? It's like getting yourself in front of people more, more frequently asking for their time, asking for a little bit of space to like show stuff to them in real time and let them know what you're working on. This is good advice for me, selfishly. Thanks, Rylan, because I'm not always great at that. I get shy. I'm like, I don't want to bother people. I don't want to take their time. They've got to be busy, right?
Yeah. I think early on, I especially noticed that too, because I wouldn't say I was the best, you know, coder in my undergrad or the most technically skilled all the time, but I had enough of the blend of the, you know, the soft skills and the hard skills to be able to, you know, really demonstrate what the data was doing, speak to it and not be uncomfortable presenting that. That was kind of where I found my niche was that I blended enough into those categories.
I like it. There's somebody at my company, Posit, who I see getting stuff done all the time and like making things happen. And I went and asked them what their secret was. And you knew what they said? I'm not afraid to bother people. I'm not afraid to inconvenience them and take their time. And I'm like, wow, I got to work on that.
All right. I saw something from Connor that I wanted to shout out in the chat to Connor says advice for pitching, buying new software, aggressively tie it to whatever big project slash issue is at hand for the company. Catching a wave is easier than creating a wave. You did that with COVID, right? Like, we have this thing. Y'all really, really need it to be real time. We could do that better if we had X tool. Thank you, Connor. I think that that is a great way of saying it easier to catch a wave than create a wave. Rachel and I find this too with community events. If you're a community builder and you can get your community event into an existing framework in your company or your community at large, whether something is already planned on like a cadence, that will be easier to do than starting something brand new, inviting everybody, creating a new time, creating a new venue, whatever it is. So I totally agree.
Catching a wave is easier than creating a wave.
Audience Q&A: Team culture and hiring
Well, I see a question from Noor in Slido. Noor, I bet you would be able to unmute and ask that live. Yeah, I see a thumbs up. All right. I'm going to highlight it in Slido for you, Noor. Hello. Can people hear me? I'm on a brand new computer, so I may be. Excellent. Yes.
Okay. What was the question? I asked so many, unfortunately. Okay. What was it like defining the culture of your data team, especially since you were there at like the very inception? So you got like founder stuff going on. And if you had to do it again, what would you do or not do? Like what are some takeaways we can take from this?
Yeah. I mean, I think I kind of alluded to it early on, but the biggest thing I looked for was just making sure we were hiring problem solvers and people that were not just pushing out the requests, but actually solving the data questions at hand and improving the process. That was always number one priority when I was looking at new hires, not any of the, you know, how good are you at writing code or, you know, what's your portfolio look like, but can I make sure that you can solve problems? I think bringing in individuals like that made it easier for me as a manager because they were, you know, actively looking through the data and asking harder questions, but also it was giving people a little bit, maybe more than they were even initially requesting, especially when someone's like, you know, I just have this Excel data that I pivot table once a month. Sure, we could automate that, but like, why are you doing that? Asking those questions. Those were always what I was really looking for.
And I think just if I could change anything, I wouldn't probably change anything in like who or how I hired, more so just trying to get those, like the connect and the tools like that earlier. Because I think we would have, we would be a lot farther in our, you know, data journey and our data maturity now, if we kind of moved to that next step a little earlier, because I think we probably had the same amount of buy-in. I was kind of waiting for a wave, so to speak, with COVID. So I think just trying to do that a little bit earlier would have been nice. Out there with your boogie board, just waiting. I'm not good enough to be a surfer, y'all. Don't don't at me. It's a boogie board for me.
Okay. Noor, that was a fantastic question. Thank you. I see Mark asked a question as well. And I'm curious about this one as well. Mark, if you were able to ask this one live, that would be amazing. Let's see if we can get you unmuted. I don't see an asterisk, so I hope. Yeah. Can you hear me? Yeah. Sorry. I would show my face, but I'm running data while talking to everyone for my work in. Hello, Ryan, or Roland. Nice to meet you and everyone. My question is, how do you determine when you're ready to move up in your company from role to role? You were an intern, now you're a manager. How do you know when you yourself are ready to pivot? Were your managers supportive? Were they adverse to you moving? I'm someone in the same boat. I'm applying for a data analyst by position internally, so I'm really intrigued on your perspective.
Moving from IC to manager
Yeah. I think it was really a pretty personal issue at the time for me. Was I ready to manage other people? I came up with a couple of them, and so moving to that management position was going to be very much a different personal experience for that. It's really just, for me, it was, am I ready to start to worry more about the people on my team than just myself and the project that I'm handled with? That was number one for me. I was very lucky, and I don't think this is the case for a lot of people, but I also had very supportive managers and directors from the beginning of my career that weren't really in the same lane as me. I reported to them, but they weren't necessarily data folks. When we started having a large enough team, they were pretty open to the idea of someone else managing that domain because I was excited about it. They were doing it out of necessity, but I was excited about it, and so I brought it up a few different times. I even went as far as making an S-bar to be a director at one time because I was just really pushing for a more formal structure where data wasn't just a subset of another team. It was its own entity and really pushing for that.
I think it's a combination of just being prepared for the interpersonal adjustment that being a people manager is and also just advocating for a dedicated group. Those are the two combinations of things that I think tell you when you're ready.
Sweet. Thank you. That's a big question. I've been fighting leadership my entire career. Like, no, no, no. I don't want to be a leader. No, no, no. But I do want a say in decisions, and I do want to influence people, and I do want all of these things that I'm like, oh, wait. Maybe I do want to be a leader. Oh, no. John on my team, I don't think he's on this, but him and I, we were in college together. We were interns around the same time, and then when I moved to a management position, that was the difficult adjustment. But wading through that was I think the hardest thing I looked at interpersonally. Obviously, you got to play it the right way, but that worked out really well, and he's still on the team, and he's one of our top achievers for sure. So that was probably what I was alluding to on the interpersonal side.
Awesome. Awesome. Thank you. That was a fantastic question, Mark, and Noor had a very, very similar question. So Noor, I will consider your question answered as well, but you had a little follow-on to that, which I'm going to ask Ryland, which was with that transition from IC to manager, what tips do you have for staying up to date technically?
Yeah. So I actually, and I saw something about marketing, but I do subscribe to this, it's called TLDR. There's a TLDR AI and a TLDR tech, these, and a TLDR data. I was just looking at my inbox now, and they just kind of give you updates on the most recent things in those spaces, and that's about the amount of content I want to consume is like a TLDR. So getting those kind of updates and then really leaning on my team that's in it every day. Luke Wagner, he's on this call, he's my senior data scientist. If there's something new in that space, I'm going to sit with him in our one-on-ones and really push him for anything new that we should be doing. And then my data engineer, Christina, same thing with that space. I'll really lean on her, like, what's the latest with Snowflake? What should we be doing? And so then I get a little bit of that just by having my one-on-ones and meeting with my team. And so that's the best way I can stay up to date because I don't really have a lot of time for R&D at this stage. So leaning on my team is that R&D.
I like that. And we also had a little reply on Noor's question from an anonymous person in a community that says, in my experience of being a new manager, you become less sharp at the day-to-day technical skills, yeah, but you have the scope to have a better high-level overview. Do you find that that's true, like better understanding of the business problems maybe and what leadership cares about or what they're worried about? Yeah, I think that's really my niche now is just taking what the business is having trouble with and knowing where and who on my team I should go to and how we can solve that problem for them. Not necessarily saying, you know, build it like this or do exactly this, but like this is the business, translating that business problem to my team in a way that makes sense to them, and then just letting them, trusting them to run with it and implement the right solution at the end of the day. That's definitely where I'm trying to go to. I still sometimes say, you know, build it like this, but I try not to anymore. I try to trust my team, so.
Most rewarding work in healthcare
Wonderful. Okay, let's see. We had an anonymous question. Oh, I love questions like this. It says, what is the work your team has accomplished that is the most rewarding? Oh, Hubert. Hubert replied and said, oops, that's me. Forgot to put my name on it. Thanks, Hubert.
You know, I think for me, and the reason I'm in healthcare is just helping clinicians, you know, taking some of that off their plate. They want to know, you know, what's going on with their patient population. They want to know ways they can improve or where the issues lie, and I think delivering a project to a clinical leader and seeing how it will improve their ability to care for patients is by far the most rewarding thing for me and why I stay in healthcare, because there's so many industries that need, you know, data experts or data scientists and people in data, and so I think that for me is by far the most rewarding. It's just helping those clinicians. I still love working with all the business folks a ton, don't get me wrong, but the clinical leaders is definitely why I stay within the health system.
delivering a project to a clinical leader and seeing how it will improve their ability to care for patients is by far the most rewarding thing for me and why I stay in healthcare
I think that's why a lot of people work in certain sectors or areas because they really care about that work. That's why I am in open source. I feel the same way, and you know what the most amazing thing about open sources is that there are all of these teams like your team that we get to enable to do all of that cool stuff, so it's just like perpetuating the goodness, yeah, across all of these health organizations, research organizations, universities, all kinds. All right, Hubert, thank you for asking that question. Sorry I missed that it was you until the end there.
Tips for Shiny apps
We have a question with a couple of replies here from Natasha. Natasha says, can you read this for me? See, I'm getting better at reading those replies first. It says, you have any tips for making shiny apps? For example, how to make them clearer or more scalable?
Yeah, I mean keeping it simple for one, and then you know reusability. So our team tries to use the same template for our apps. That way it's both consistent for our end users, but also we're not reinventing the wheel every time we're building a shiny app. And so I think that was the biggest thing early on is just, you know, keeping it simple and then also iterating. You know, we try not to... Deliver, you know, there's a ton of different phrases for this like drain the ocean We try not to give everything all at once we try and start with like a few Important KPIs or metrics that the end-users know they want And then also just like a way to see the raw data and then we just make sure we iterate with those And we don't we don't ever want to hand something off and then say see it never We want to make sure that we're saying like let's let's circle back in a month or or so when you've had time to really Get hands-on with this first iteration and then let's see where we can improve it and let's make sure that we're staying very hands-on with it so
Fantastic. All right. Thank you, Natasha.
Impactful data products in healthcare
We have a question from Ken Ken Roosevelt if you happen to be available to ask your question live. I would love that. Well, let's see if you're available. Or if you forgot to put an asterisk on there he is. Hey, can it can you hear me? Yes. Hi, I am also. Hi Ryland. I'm also in hospital and healthcare Data, so my question for you is what is a data product like a dashboard automation predictive model, whatever? That your team has developed and you feel has made a substantial difference for the care of your patients or hospital operations And on top of that, If you are thinking of a few, Is there? Maybe something more hospitals should emulate and bring to their own health systems that you can think of
Yeah, so the one that comes to mind initially we did we implemented probably Seven or eight years ago, and it was pretty low-hanging fruit now when we think about it, but we have a pretty large labor and delivery like women's hospital here and They really didn't have a way to understand what the birth numbers were gonna look like they were growing almost exponentially at the time Not really just like a linear trend. They were really booming at the time and every month the month It seemed like they never knew how to staff effectively. And so we just put together a pretty simple Model for birth birth numbers using almost exclusively just like patient data and delivery data
To just show them like here's the next six months like trying staff like that something that doesn't seem super novel right now But you know, it was a big value add for that women's hospital group and it's something they're still looking at to this day
one of my senior data scientists actually was just implementing an update to that to try and Trend a little bit like get that a little bit more more accurate. But that's probably the one that jumps on my mind immediately
With that are you looking at like pre bookings like Because you know, like they're expected delivery dates Yeah, so we mostly are using the the historic data, but we do leverage the if so the exact terminology would be Like an induction a scheduled induct like we have visibility to that so that we can we can see that not necessarily all the Spontaneous deliveries, but if there's a scheduled induction, that is something we incorporate so that we have a little bit of foresight
And it's not just here's the last six months. What's it going to be? You know next month, but we try and leverage that data and there's other measures that we're looking to add Now that we have a little bit more expertise in that magic matter to like Like the other hospitals in our area, you know Once we're able to start getting that with our state Collaborative then we want to start looking at that and see if their numbers are going up and ours might trend with that or if That's taking away. We'll be able to add that one more factor to that analysis, too
Thank you. Thanks Ken for the question Don't feel like you have to stop at one I like good ideas, you know, I'll steal as many as I can
We all will well. Oh go ahead Ryland. Well, I was just I was just thinking Like the other one in our finance group that's been asked for a long time is we have like our daily revenue That's just a snapshot today and and historically and I know It's something Luke and I have talked about to help them Our CFO is really data driven and something to help them You know project out a little bit more from just our cost accounting system is is using but that's not something we've implemented yet But I know it's something that's been talked about quite a bit within with our CFO is just you know Getting ahead of the expected revenue and not just seeing it, you know yesterday and month in the month today year to date But seeing what it could look like and what it's going to look like in different scenarios
AI and automation in the team
Awesome. Well speaking of stealing good ideas I noticed that in the Q&A nor had asked a question that was to everybody So I just put it in as a free text poll if you are on Slido and you're in the Q&A section You should be able to slide the little slider at the top and go to polls and answer this I'd love to hear that question was how do you find a mentor? How do you identify pursue or find a mentor? And I think I believe that when you put stuff in you should be able to see everybody's answers Let me know in the chat if that doesn't happen and I will try to make sure that it does happen But I'm hoping that it does it says poll results will be visible, okay
Thank you, Nora for asking that question. I appreciate it. Let's hope we get some cool answers There was an anonymous one in this in the Slido So I will read it out and it says if you were asked to go into a team that wasn't using AI much yet And your goal was to have a conversation about where AI could add value. How would you approach that? What areas or questions would you want to explore with that team?
Yeah, so I wouldn't say my team currently uses a lot of it I mean, I think we're it's kind of cost prohibitive for us as a healthcare system to have like frontier models You know being code a sit like things like that and so I think we we're kind of a You know a case study on that already in that we try and just find the the things we don't enjoy doing so for me, that's you know spinning up the Like if we do have a need for a custom changes to a shiny temp or a shiny application I don't necessarily want to go in there and fix every you know picker input and Every you know high chart that we use like I I'll use AI a lot to just say like this is what it is now I want to switch this stuff and then it will just kind of Like tweak that code enough for me that I'm not You know in there deleting rows of data and probably breaking stuff at this point. So
I think for us it's really just a matter of like Finding what you don't enjoy or what takes you the most time and then automating that piece of it But I think it's going to be different for everyone and what I what I would say the biggest you know issue that I see is Treating it like a you know Everybody should be using it the same way because I would hate for to tell someone on my team To use AI to generate their their code if that's what they enjoy the most about their job So then I would maybe say well find something that you really like and then you know So then I would maybe say well find something that you really don't enjoy is it documentation, which I don't know anybody that likes documentation, but Using it to document the code and like what you're trying to do putting it in your github Read me whatever that may be finding that issue or that pain point and that's what you should focus on using AI for
Finding what you don't enjoy or what takes you the most time and then automating that piece of it But I think it's going to be different for everyone
Right because you could just write your own code and then have it write your commit message for you Yeah, i'm terrible at writing messages I usually I'll take a stab at some Yeah, I usually take a stab at some like if I have a concept for something like recently We have been working with our surgery department on block utilization and so That can be cumbersome when there's multiple surgeons in multiple different rooms like who's utilizing what so I I took the initial You know I wrote the code that I thought made the most sense and then I gave that to AI and I said am I doing this like Can you notice any optimizations or ways that I can make this run faster, etc and then that's how I had it tweaked because again, I don't know at this point in my Like i'm not up to date on the current best practices for every way of writing code And and I also don't want to ask my team 30 questions a day about like oh should I be using this function or that? And so that's where I've I found it useful when I am, you know in the code writing it
but I also use AI mostly for like presentations and uh, you know writing up documents or our governance charters things like that that are What I don't enjoy is the monotonous, you know management work, but Yeah, go figure out what you don't like to do. I like that. Awesome. Thank you. Ryland
Um, we had a quick one that we can get out of the way, which was anonymous It says is your position in your health system in operations or is it in academic research? I think it sounds more Operational we're in it. My team is in it. Um But we do work with um, especially luke who's on this call. I've alluded he he works in the research space a lot more than Myself we help with a lot of our, you know, some of our surgeons and physicians Uh, they still they like to submit papers and do research Uh, he does a lot of like the statistics and data work for those individuals and so You know, we are technically in it, but we really hit as many Domains as we can we work with our college. We have a foundation So it's kind of just like, you know anywhere that anybody that needs data help. They just try and come to us
Deployed tools and day-to-day applications
Awesome is there an example? um Sort of on this this line of like, where do you sit? Do you have an example that you could talk about? that's not super super private, I guess of like something that you do have deployed currently that people use day to day or that like comes back to you as um something that is Being constantly iterated over time. Maybe every quarter gets changed and people are using it a lot
Uh, yeah, so there's one that comes to mind so we have a board statistics um Project I would even say I'd call it a dashboard, but it's kind of a little it's more of like an application um, it used to be um a big a large process that finance and our uh, Him group would work together on to get like, you know, all these different pieces how many patients discharged births surgeries And that's probably our one of our most viewed tools at this point Um, and it really wasn't even it wasn't a lot of uh work it was just kind of identifying where it all comes from and then setting up automations to make it a pretty General shiny app that had a bunch of you know Sidebar options menu items that they could click through and find all sorts of information and it works from you know Our ceo can get a high level summary and then our finance team can drill in to look at really specific information So it kind of works for you know, any level of user
Um that one and then we have another one, uh It's called it's our birth certificates application and it sounds odd, but it was something that our birth, uh, Clerks, so when a baby was born they were using an access database to plug in They'd go into the chart find the baby that was born plug in their information print it out hand it to You know the parents and so we built a an application that kind of automates all that pulls the data from the emr Straight into a shiny app The birth clerk can just click that baby that they see was born And then that auto fills what they're taking out and so it saved a lot of time That's probably actually our most used application because we have you know, 500 600 babies born a month in our health system. And so yeah, there's you know, 20 20 babies a day that need a birth certificate, so
And it's more like a it's more like collectors or like not a not the technical birth certificate obviously But it's like what we sent home with the parents. So Yeah, yeah, like the the not the not the nominal. What do you say? Yeah, like a keepsake Yeah, it's like their keepsake more than anything. Yes, but it was something that we they pride themselves on and so you know, we we gave that to one of the hospitals and then another hospital wanted to start using it and um, you know that one that one probably gets the most clicks per day for sure, but
I think that's a really good example to sort of tie into the question about What problems do you talk with people about in their department for solving with ai like that same question applies to just solving with automation with shiny apps with Quarto dashboards with with tools at data tools at all, right? And so this little problem that feels really small and for you you're like, oh, yeah, we can solve that for you That's actually a really big problem for five to six hundred births per month, right for these people who are doing it manually all the time They might be like this is just what it is. I just do this thing manually, but they don't know what's possible So you come in with this? I know what's possible mindset And there you can blow some folks minds. Yeah, that's amazing
They might be like this is just what it is. I just do this thing manually, but they don't know what's possible So you come in with this? I know what's possible mindset And there you can blow some folks minds.
I think the biggest thing to be careful with that and this kind of speaks back to early on like how do you get wins? Is when you do something like that, you never want to phrase it as you know, i'm automating your job It's always supposed to be You know i'm automating this piece of work kind of like what ai is going to free you i'm going to automate this so that Frees you up to do things you actually enjoy doing because I know these clerks didn't enjoy you know typing into a microsoft access database and then printing like they actually like the interaction with the parents and the babies and You know providing all that in a timely manner and so it's really important to when you do automation projects and things of that nature to make sure you uh, you know propose it as a Value add not a replacement of someone's, you know job duties. So
For sure Yeah, because that thing that someone has to do manually to go type that in and get that out It's probably not something that they are in super enthusiastic about not often though
Finding a mentor and peer learning
All right. Thank you. Rylan. I think that was a fantastic example. Um I wanted to point out that there are some great answers on nor's question that was in the poll I have closed that and I opened another one so that we can get even more answers to the question that just got asked Which was like hey if you go into a team that wasn't using ai I also put in there or automation Much yet and your goal was to talk about it. What would you ask them? So if you have advice on that you can go pop that in there And then if you would like to see what everybody else answered about how to find a mentor go. Um Oh, actually, I don't know if you can see it now that i've closed it, but I will I will mention a couple of these, uh really quickly because christy had answered keep in mind You might need a different mentor for different pieces of your skill set and I totally agree with that Um, I tend to have only found mentors through academia slash work So they were right there, which doesn't really help you but also here at the hangouts or the lab or the discord Those are great places agree christy. We have a whole discord We have this wonderful community here where there's 107 people in this room, right and we're all in sort of the data space um connecting here Building relationships go connect on linkedin You know, if somebody seems really nice ask them. Hey, do you want to have a zoom chat sometime? I'm I'm interested in what you do or there's something that you would like to you know Get a little bit more information about you don't have to immediately ask someone. Hey, can I have a job right? Like you can just have a conversation with somebody rylan Do you have any extra tidbits of advice for mentorship or finding a mentor?
You know what I was thinking was kind of the first comment you said was that um, you know Don't necessarily look for a singular mentor try and find, you know different pieces of Mentorship from as many people as you can and also don't look past, you know like your the colleagues people in your same, you know tier of the organizational chart because You know, I I learned from all of my team and when I was an individual contributor I learned from our other individual contributor a lot We had complementary skill sets. And so I learned a lot from him. Um, maybe hopefully he learned something from me and so I think don't look pat don't always look to you know management or hire for mentorship try and You know look at the people to your left and right as far as what you can learn from them, too
Absolutely. Also if you are working with interns, I have learned so much from interns like people who are either younger than me or just Behind me on the the ladder of experience I've learned so so so much from those people. I think reverse mentorship is absolutely a thing So look all around you in 360 degrees Including to other industries because the best thing about the hangout for me is that we're all from different Industries and while we're learning from Rylan who's in a health care system You might be hearing things that Rylan's talking about that apply directly to you in auto insurance property and casualty insurance In research in some other field so don't stick to just your industry either
Okay, we have 10 minutes left. Let's go grab some more questions And don't forget there's a question in the poll if you would like to go answer it Um, let's see. I saw one from Morgana And I think we have morgana live here. Hey morgana. Would you like to ask your question live? I feel like we asked this question slightly already, but I but I i'm going to allow you to elaborate on it if you would like because there was a um sort of Add on to it. Oh, but I can't hear you
Oh, man All right, I can't hear you I could ask it for you. Um, but it was How do you balance trying to become a leader in a more strategic role with key people? More strategic role with keeping up with your more technical skills
yeah, I wanted to ask this one because I feel like There's more emphasis on the strategy part here than just the technical part Like how do you do the strategic learning as well inside your organization?
Yeah, well, I I do think it's a give and a take. I mean if you want to balance multiple I mean some people can work, you know after they're nine to five They'll work into the night and I think that's great for people that can do that Like I have you know a three-year-old and a three-month-old so i'm not really sacrificing my evenings anymore for To to continue to upskill technically so I do think it is somewhat a give and a take But the strategic thing I think really just comes from like, you know I i've been lucky to have some great directors and so grabbing a little bit of what each of them has done Well, and then trying to just kind of pocket that information for my own skill set I think has been the most valuable for me personally
Um, but you do a lot of it is just learning on the job I don't I don't think anybody's ever a hundred percent ready to move into a strategic role before they've done that And so a lot of it is just you know, learning as you go um making sure that you um, you know learn from any mistakes you make because you're gonna I make mistakes and everybody does so I think it's just kind of a combination of Uh, you know learning from the mistakes you do make and learning from those around you and those similar rules
Career advice
Fantastic, thank you so much morgana for your question Um, we have about eight minutes left and I think we can get one more side of a question in but I would also like to ask you a question that we try to ask everybody which is uh, do you happen to have a piece of career advice that You wish you knew before you got started or that you find helpful now, or you'd like to share with others
Um, you know, I I think uh, it might be cliche but like networking and you know Making good impressions is very valuable. Uh, you know, you don't you're not going to get super far if people don't like working with you I think and so I think being making sure that you're uh, you know As cordial as you can and helping deliver on things and you know Sometimes you have to hear out crazy ideas or things that don't you know, don't really make sense But it doesn't really do you anything to say? Well that that's a dumb idea or that won't that won't work. So I think just making sure you know when you you hear uh someone's idea trying to interpret that into a way that Uh, you know, you can deliver while doing it the right way
and then also just always making sure you're speaking to your audience and uh, you know not saying a bunch of technical jargon to a nurse or Uh, you know saying a lot of clinical things to an it person So I think really the combination of those two things I think has gotten me a lot farther. Um than if I did the inverse
So I think really the combination of those two things I think has gotten me a lot farther. Um than if I did the inverse
Prioritizing projects with limited resources
Perfect, I think that's great. Thank you. Rylan Well Hubert had a question and this time Hubert's name on his is on it so I can call him to ask it live Um, and we've got about six minutes left. I think we can do it
Oh, no, we can't hear you. Oh, no, he's here. Okay. Sorry. There was a delay. We got you. Thank you. Thank you. Rylan and We should do coffee one day. We're since we're both in omaha for sure. Yeah for sure
My question is how does lack of resources impact its scope of your projects? I think in a group like yours who has a reputation of getting things done and producing things that people use and like Uh, you've all you how do you pick which projects get done when there's always too many on the board to do too many requests So, how do you prioritize? Yeah, so I think um for better for worse like financial impact does kind of jump the the queue Uh, you know when there's uh, our cfo comes to us with a project that might you know have downstream financial impact that's gonna You know, obviously give us the most wins um and Similarly, you know if there's if it touches executive leadership Then we often will jump a little bit into those projects before we do something that maybe Uh is just for you know, general data knowledge or data analysis
and then we also have a lot of you know, cms and state bodies that Require certain things so those kind of automatically get categorized a little bit higher but then you know of all of those are not if none of those are true, you know, then we'll Uh, then then it really becomes a judgment call to see you know Like where do I think that this data will have the most impact how many is it going to touch the most patients? You know, we have a lot of niche groups We have a lot of very large groups so you know if someone from our like our women's hospital comes to me I know there's going to be a larger impact than someone from you know, like uh Our one of our more remote hospitals that gets much slower throughput. So I often will just prioritize it on Which one's going to touch the most patients and or the most people individual individually, so
Vendor agnosticism and EMR
Perfect. Thank you. Thank you for the question hubert I wanted to call out a few answers to the open poll about what would you talk with people about or ask them about? Um, if you were getting them to add automation and stuff one was Um that you know ask them what they spend the most time on or what they find least enjoyable But I really like the little follow-on bit which was hey, it's really hard to get people to adapt to new processes But if you immediately make their job easier or more enjoyable Adoption will be easier. Yes. That's that like connor said easier to catch a wave than to um, create one right or or build one um same there's probably a wave already of Maybe unhappiness or discomfort that you could ride
All right Let's see I would also really really like to see if we can grab one more I think we had a small one because I know that ken had asked. Hey, I forgot to ask what emr Are you on emr is electronic medical record if you feel like sharing that you could Forget it. We use we use cerner
Cerner, okay. We had a few different people who use different things on the hangout, which has been really really interesting to hear and Yeah, the only thing I would sorry to interrupt you The only thing I was going to add is my goal is to be as agnostic vendor wise as humanly possible So to me, it doesn't necessarily obviously there's you know Technical differences between how you get the data But like my goal as a leader in data is to make it so that we could swap out Of any vendor and it doesn't destroy things downstream um, and so You know like we do use cerner but my goal is to if we switch to epic Be able to or any of the other emr be able to you know, have that not Uh destroy things downstream similarly with any of our other vendors We just switched from chronos to kujenda. The goal was to make sure all of those reports stayed the same Um, you know when we did that switch so it doesn't uh impact us a significant amount Which emr we use but that's just something I like to always bring vendor agnostic is one of my favorite terms I use so
The only thing I was going to add is my goal is to be as agnostic vendor wise as humanly possible But like my goal as a leader in data is to make it so that we could swap out Of any vendor and it doesn't destroy things downstream
Nice. I've never heard anyone talk about that before. I have heard about uh From different hangout people about epic. So that's interesting. It seems like epic is like the big player. Yeah, they're probably the bigger Yeah, they probably the most market share I guess Yeah, you have some agreement in the chat as well as like vendor lock-in is a dangerous thing Yeah vendor lock-in is yeah the same term basically. I just yeah sounds like a great a great piece of advice
Closing remarks
Well, awesome I would love it if everybody who is at the hangout now could tell somebody about the hangout That would be great. We've talked a lot about like getting in front of people networking um This is such a great place to meet people I'm, so excited that hubert and rylan mike will have coffee. I really want more of that for more people at the hangout. So
If there's somebody that you feel like would benefit from this Little bit of data community analysts data scientists data engineers data leaders, whoever it is Invite them to the hangout. We don't really have a super broad reach right like on linkedin and I'm often in just linkedin jail. I post on linkedin and like two people see it so Share the hangout with somebody today. I would love that and then also please register for virtual for the posit cov um coming up in the middle of september september 14th through 16th because I want to hang out with you on the discord server And isabella too and rachel is going to be in person if you're making your way to houston Make sure you give rachel a high five. Okay
Rylan, thank you so much for hanging out with us today sharing all of your knowledge all of your examples This was so much fun. I hope you had a good time, too Yeah, thanks for having me Fantastic and thank you to anyone from rylan's team who joined and hung out with us I do think that we had somebody on here luke for a little while maybe halfway through So thank you to luke rylan. Please extend that. Thank you to luke Everybody is saying thank you to rylan in the chat amazing. We will see you very soon next week everybody We will not see you on tuesday at the data science lab because we don't have one Um, but if you would like to save the chat because there were cool links Um saved or shared in the chat You can click the three dots in the top right hand corner of your chat screen and save the chat
um If you made your way to the discord that is another place where you can ask like hey something was was shared at the hangout And I would like to see it absolutely ask there We will put the chat the file to the chat either isabella or I in the data science hangout channel on the discord All right, everybody We love seeing you. Stay safe. We'll see you next time. Bye