Resources
Dec 28, 2023 |
59 min
|
470 views

Sean Nguyen @ S2G Ventures | Data Science Hangout

We were recently joined by Sean Nguyen, Senior Staff Data Scientist at S2G Ventures to chat about guiding newcomers into the data science field and promoting business integration of various data science tools like Posit Connect.

Featured Leader Bio:
Sean Nguyen is a Senior Staff Data Scientist at S2G Ventures, where he works on the data science and technology team. He focuses on developing data products and identifying actionable insights for the firm and portfolio companies. Before S2G, Sean was a data science fellow at Insight Data Science where he developed a machine learning model to predict the outcome of trademark infringement lawsuits.

Sean received his B.S. from The University of Michigan-Dearborn and his Ph.D. in Cell & Molecular Biology and Environmental Toxicology from Michigan State University. During his time at MSU, he worked as a data scientist in the graduate school and as a commercialization intern in the technology transfer office.

________________________

ā–ŗ Subscribe to Our Channel Here: https://bit.ly/2TzgcOu

Follow Us Here:
Website: https://www.posit.co
LinkedIn: https://www.linkedin.com/company/posit-software
Twitter: https://twitter.com/posit_pbc

To join future data science hangouts, add to your calendar here: pos.it/dsh (All are welcome! We'd love to see you!)

Thanks for hanging out with us! šŸ’›

image: thumbnail.jpg

Transcript#

This transcript was generated automatically and may contain errors.

Hey everybody, welcome to the Data Science Hangout and our last Data Science Hangout of 2023. I know people always say this, but I seriously can't believe that we're already here at the end of the year. I want to say a huge thank you to the 47 leaders who have joined us at the Data Science Hangout this year. Thank you, Rachel, for hosting. And those three sessions where we all were the featured leaders. And who knows if this is accurate with Zoom, but the over 2,600 people who have joined us at Data Science Hangout since we've started them and the core group here who keeps coming back and hanging out with us each week. So thank you all so much for making this space what it is. I really love getting to spend time with you all.

If this is your first time joining us today, so nice to meet you. I'm Rachel. I lead customer marketing at Posit. If it is your first hangout, say hi to us in the chat because we'd love to welcome you in as well. This is our open space to chat about data science leadership, questions you're facing, and getting to hear about what's going on in the world of data across different industries. So we're here together every Thursday at the same time, same place. We'll be on a bit of a holiday break here for the next few weeks, but we'll be back on January 11th. So if you're watching this recording on YouTube later and want to add it to your calendar, it will be in the details below on how you can join us live.

We're all dedicated to making this a welcoming environment for everybody. So we love hearing from everyone, no matter your years of experience, titles, languages you work in, or industries. And it's totally okay to just listen and hear if you're out walking your dog, if you're in a coffee shop or something. But there's always three ways you can jump in and ask questions or provide your own perspective on certain topics. So you can raise your hand on Zoom and I'll keep my eye out. You can put questions in the Zoom chat and just put a little star next to it if it's something you want me to read out loud instead. And then we also have a Slido link where you can ask questions anonymously. And we'll share that in the chat here in just a second.

And I love Mike's suggestion in the chat here to get everybody going because we love the conversation that happens in the chat. But share your favorite holiday film in there. I'm taking Mike's suggestion from their Pfizer hangout too. But with that, I'm so excited to be joined by my co-host today, Sean Nguyen, Senior Staff Data Scientist at S2G Ventures. And Sean, I'd love to have you jump in here and introduce yourself and share a little bit about your role, but also something you like to do outside of work too.

Yeah, so I'm Sean. I'm a first data science hire at S2G Ventures. We're a venture capital-based company based in Chicago. We're investing in food and oceans and clean energy companies. And so we were founded in 2017. And I was kind of brought on in 2021 as the first data science hire to kind of flesh out the capabilities of the organization. And my background is in molecular biology. And so I was really interested in working in data. And then I kind of learned about this thing called data science. And then I got really addicted to learning how to program. And then I kind of took those skills and took it to S2G. And then in terms of things that I like to do outside of work, I really like photography and cooking. So I'm a bit of a person that likes to try out new restaurants in Chicago and cook at home.

What's one of your favorite things to cook?

I like to cook different cuisines. I don't know. It depends on what I'm feeling. So sometimes I'll create Thai curry paste from scratch and just kind of make that and use it for different things. So I try different cuisines.

The proof of concept and getting buy-in

Well, I get to start off with some of the questions here as we wait for them to come in from everybody. But I just want to say I loved your composite comp talk from this year and was just listening to it a few days ago. But I love when you spoke a bit about the POC or proof of concept that you did. And I thought it might be helpful to start by talking a little bit about that so we can all get some context into some of the work that you do. Would you be able to explain a bit about the project that you came up with to use for your POC and kind of get buy in across the company for your project?

Yeah. So as the first data science hire within the org, I knew we wanted to get information to our business units. And so what I wanted to do was run analyses and send it to, let's say, my manager. Right. The thing is, is that when I created our markdown document at the time, in order to share it with with her, I had to basically render the HTML file and send it over. Right. Because I didn't have like a server. Right. Because this proprietary information, we can't share it easily because you don't want it out in the world. So I'd send the email attachment and send it to her. And then she she would be like, oh, I this is great, but I don't want to download something and have to open it. Right. So I want to be able to click a URL link.

And so in terms of what it was, what it was, was it's kind of an analysis of anything that she wanted to know. That's like perfect for for for her job. Right. So let's say like here's the meetings that you have during the week. Right. And that's really insightful information that you can't necessarily get from, let's say, a. Like a dashboard or something that is super specific to them. Right. So like in terms of the idea of the POCs, that is to kind of almost give that nugget of like, oh, wow, this is really useful to me. And then being able to surface that to them. Right. And then what that does is it gets that buy in, because oftentimes when you're trying to, let's say, get a new service or a new product. Right. Some people will look at the sticker price price, like a sticker. It's like, oh, wow, there's no way I'm going to pay that. We already have Tableau. We already have Power BI, whatever. Why do I want to get another, you know, data science product I've never heard of? So in order to kind of almost get the buy in is you kind of give them that like, oh, wow, sure, the art of possible right to be like, oh, you know, here's this cool thing I was able to do. And then they're like, oh, my gosh, I really, really like that. Tell me more. And then that's how you can kind of get them to be like, oh, this is really useful because you kind of are servicing information that's really relevant to them at that time. Right.

Communicating insights proactively

Yeah. And I think that's something that's come up quite a bit on the hangouts, too, which would be good to chat a little bit more about is the communication piece of your insights, but also. Realizing you might need to have some different approaches like not everybody's going to go and look at a dashboard, they might need a summary of insights in an email. And I'm wondering what lessons have you learned about that communication piece?

Yeah, so I think the idea of being able to proactively find information that might be relevant to people, because like to your point, like I can create 10 dashboards and then just check the dashboards. Other people, they don't have time to do that. Right. And so sometimes you can be like, oh, it looks like you're meeting with, you know, Mary from whatever company. Did you know that X, Y, Z thing happened? Right. And then that then they can go armed, ready with this knowledge whenever they meet with this individual and they don't have to check a dashboard. Right. And so being able to find whatever information, whether it be from a different data source, some analysis or report, and then you're able to kind of send them an email. Right. Or that or some type of alert dynamically. Right. So I use the Blastula package with an hour to kind of send emails programmatically to people. Right. And so that's how I was able to get them to to really like latch onto that, because to your point, you know, it's it's people don't want to necessarily have to check something. Right. It'd be nice if it was delivered to you in your inbox and you can be like, oh, wow, this is really helpful. And then you go on your way.

It's shared into the chat, the example that you referenced in your talk. But Kelly O'Brien at Posit built this example where she had like a cookie cutter Etsy shop and she showed how she used Blastula and Posit Connect to send out these conditional emails. It's just a fun example I wanted to share in the chat. Oh, yeah, it's great.

And I'm seeing a lot of love for Blastula here in the chat as well. So feel free to share some. Oh, I think you already are. Share some resources on that, too. Thanks, Marcos.

George, I see you asked a question here in the chat. Do you want to jump in? I know it's a question kind of for the chat as well, but I thought it would be good to discuss live, too. Sure. Yeah. Thanks, Rachel. Yeah, that got me thinking about just prior prior role that I had, you know, and I was wondering if anyone else had a prior job in sales. I think the way you put it, Sean, where you said, you know, the person might ask, the person might ask, you know, I already have this. I already have that. Why do I need this new thing? Like, that's exactly like the kind of sales question and like interaction that I used to have in a prior role. But I never really thought about bringing those skills here because I was never really good at it. It's not like my my natural state to try and sell things. So just wondering if there's any best practices that maybe people have in general from being in a sales role. Maybe you sold cell phones. Maybe you sold cars. I don't know. It seems like it's very important in what we do.

Yeah. I mean, I don't have experience in sales, but I think to your point, I think it's about trying to find that the way that you can provide value. I feel like I'm reiterating my talk. Sorry. So it's to find value that that the person would be interested in. Right. And so then you're trying to make their life a little bit easier. Right. Whether it be selling something or you're trying to give them information, in my case, information that allows them to act and, you know, kind of do their job a little bit better without friction. Right. And so at least for me, it was just building different different reports. Right. That allows them to to to do whatever whatever they need.

One thing I did recently, I think on Tuesday. So we do. I decided last year to make a Quarto reveal JS document. So a slide deck of like we call it S2G wrapped like the Spotify wrapped where I summarize all the things that we've done in the organization. And it's like more than just a PowerPoint deck because it actually embeds all the different widgets and dashboards. And so I use like GT with the interactive interactivity with it. I used all these different charts from like higher charter and all that and allows people to kind of interact with stuff and kind of see everything as a summary. Right. And it's one of those things that I got so much good feedback. Oh, my gosh, this is amazing. Right. And then allows allows the folks to to kind of see like all the work that we've done. And so that's like one thing that I kind of created. And I would encourage other people to do that is to use like Quarto. Right. You can create an HTML file, a PDF, or you use like the reveal JS presentation or you can render it as a PowerPoint as well. But it allows you to go from just routine analysis to something that's really amazing that you can share with people in your community.

I love that idea, George, about like using that prior experience that people may have learned in sales. And it was just making me think. I mean, I was in sales at Posit before, but like one of my first jobs was actually working a beer cart on the golf course and like driving around in the golf cart. And that in a way was also a sales job, figuring out different ways that you could like go out of your way to create a better customer experience. Like if somebody mentioned they like needed sunscreen, like going to grab that and like bringing it out to them. And so it was just kind of like making me think about like when you're serving these insights as well and like figuring out the best way to share that information. So your end user has the best experience possible. I always love being able to tie in like past roles like that and lessons learned. But RM, I saw you had a comment on that thread too. So I was curious if you wanted to add anything on that.

Hi, thanks. No, not a whole lot to add, but just want to say that sales skills come up in surprising ways. You know, you might have the most wonderful piece of analysis, but you can't convey that to your stakeholders. That's essentially worthless. So it's important to remember, no matter how cool your code is or how powerful your analysis is, you still have to sell the thing. I know that sales sometimes is downplayed a little bit, but it's an important skill to have.

No matter how cool your code is or how powerful your analysis is, you still have to sell the thing.

Absolutely. Yeah, for sure. I've noticed that firsthand, right? Where, yeah, you can have the

fanciest models or anything like that, but unless you can surface it to them and something that's meaningful to them, it's exactly that point, right? Where you can have, if it lives, was it, I think it was David Robinson, right? If it lives on your laptop, it's essentially useless, right? You need to ship it or send out something so that people can actually see it or appreciate it, right? Because like, yeah, it's great on your computer, on the code, but like, will people actually be able to use it, right? It's cool that you know about it, but you want to share it with everyone else, right? So, yeah. Absolutely. Alan, I see you had a question. You want to jump in here?

Yeah. Hey, everybody. Sean, I'm really curious about how you figured out, as the first data role in the org, how did you and your leaders and your peers figure out what success would look like for you? Because on one hand, you've got some autonomy, you can do PSEs. At some point, you need to figure out, like, does that get formalized into, you know, like, am I successful at my job? And I'm wondering how that, what that process was like for you. How did you, how did you get agreement on like, okay, I think this is what I do now, and people are, you know, happy with that. What was that process like? How did you figure that out organizationally?

So, I think my role has evolved over time at S2G, right? And at first, it's just that we were trying to, we had all this information, right? We wanted to be able to, because like at the end of the day, technically, for the bottom line for S2G is like, being able to invest in innovative companies, and then hopefully being able to generate market returns for the fund, right? At the end of the day. And so, it was, for us, it was clear as like, hey, are we driving value for our portfolio companies? Because if they're successful, then in the end, we'll be successful. And so, it was clear for us as being able to identify, let's say, investors that would, we would want to co-invest with or other companies, right? So, it was clear from there, like, if I'm able to identify investors or companies that are relevant to us, then therefore, it will help the bottom line of the organization.

And then, and so, in terms of like, how I'm successful, so if I'm able to identify or create different models that allows us to, you know, find out a company that's a diamond in the rough, or an investor, or someone that we may not have interacted otherwise, right? We allow them to find them by doing analyses and things like that. And therefore, that's how, like, at least for me, I was successful, right? And so, one way to do that is to be able to kind of interview the different stakeholders, not necessarily your manager per se, right? You get guidance from them, but then you interview stakeholders, identify their pain points, right? It's like, oh, I, every time, you know, I'm scrambling right before my meeting to find the talking points to whoever I'm meeting with, right? So, then I was like, okay, let's take that, take the different data sources using the contextual knowledge of, like, the different data sources that we have. So, for us, we use Salesforce, other data sources, right, trying to integrate them, and then, you know, provide a proof of concept or whatever, and then you drive value there, because then you're able to say, like, remember when you went to me and you said, I have trouble figuring out what I'm going to talk to about with this person. I'm able to provide that. Therefore, I give them something that they like, and that's how you can kind of be successful in that regard, at least for me, right?

Just to follow up a tiny bit, was there anybody who, again, that's the first role of the kind, was there anybody who thought you were going to be doing something different, and you had to, like, reconcile, like, oh, no, I think it's this, and figure it out together? For sure, yeah, I think, yeah, we had different tools, and I think for me, I knew what I wanted to do, right, but then sometimes you can't do that, because you need, like, the data foundation there, right? You have duplicate accounts, you have, like, there's so much, like, things that you need to cleanse and to be able to run the cool, fancy models, which I would love to do, but the thing is, is, A, sometimes the data's not there, it's not clean, so it's, like, getting, I think, for me, is, like, you create the POC, and then you, because you know what you need to do in terms of cleaning the data, get the data sources, and actually make it so that it's usable, so that you can do it, relevant analysis, right, so you kind of give them a nugget that they would like, and then on the back end, you kind of fix stuff, and make it cleaner, so that you can actually make, you know, more feature-rich POCs, but yeah, definitely, it's a struggle, right, A, getting the buy-in to get the tools to be able to do the thing that they want, and then kind of getting the buy-in from them, but yeah, it's definitely a struggle, so, but then you have autonomy, I had a lot of autonomy, right, so that's, like, the pros and cons of, like, being on a big team versus, you know, being the solo contributor, yeah, nice.

Working alongside BI tools

I know you mentioned in one of the answers before, like, trying to advocate for tools when you already have something like a BI tool, but I was wondering, do you also work alongside with the BI tools, or are there any use cases with that? Yeah, so what I've done recently, so I, we had some BI tools before, and, but then what we didn't have was, because sometimes those tools require, essentially, more or less model data layers, where it's, like, pretty clear, like, when you're querying from this, you feel confident in the data that you're pulling from, in terms of the rows, and things like that, but what I had to do is, I had to create ETL pipelines, so it's just getting data from our raw data sources, transforming them, and modeling them so that they're actually usable, and then that actually feeds our BI tools downstream, right, and so, at least for me, we use Posit Connect to have scripts that take the data, do the thing, and then send it back to our data warehouse, right, and then from there, our BI analysts can take the data and do whatever they need to do, and feel confident about it, rather than pulling from the raw logs, right, we're able to kind of just point to, they call it a model data layer, right, and so then you know confidently, with high confidence, that what you're displaying is accurate, right, so it's not like we're trying to replace one another, we kind of live in harmony, right, so I'm able to take the data, clean it, and then run my own analysis, but then other people, like other not, power users that can use it to run analysis themselves, and not have to worry about is this correct or not, so yeah.

Attributing success: modeling vs. delivery

Thank you. Frank, I see you had a question a bit earlier, want to jump in here? Hey, Sean, thanks for all your insights so far. You started talking about modeling, and you mentioned, right, like building a model and finding the gems that wouldn't be seen otherwise, and it got me wondering, a lot of this conversation has been focused on sales, or right, like how you pitch data science, and the value, and how you get information to the right people, at the right place, at the right time. I, the question popped into my head, if you had to ballpark it, right, this is not a super quantitative answer I'm looking for, but I'm curious, if you look back, what would you say you attribute your success to, right, like 50, 50 percent, is it that the sales, and the operations, and how you, how you delivered data science to your, to your organization, versus the modeling, and the statistical rigor that you do, kind of like behind the scenes?

In terms of like how, like where do I spend my time, or like in terms of what metrics, sorry? Yeah, no, like what do you attribute your success to? Like it sounds like you've been pretty successful in your role, so if you have to look back, and reflect, and say like, oh my success, right, like I do a lot of like delivery, and like selling this to the organization, yeah, yeah. Yeah, so I would say it's, it's, it was a struggle at the beginning, right, like there's no doubt about it, of getting all the, A, the data, getting the model, and then we like, what I did was, I was able to, I would say maybe spent like, I don't know, 75 percent of the time getting the, the value, in the sense that like getting the nugget of information, but it was in like a, a Colab notebook, right, so I had this machine learning pipeline, and it was like essentially a white glove service, right, of like, Sean, give me the top, you know, 100 investors for this company, or whatever, right, and so I would do the analysis, and run that stuff, but then they were like, oh, you know what, I want someone that's going to write a 10 million dollar check, so then I have to go in and change it, change it, change it, so then what I ended up doing was like, okay, this is not scalable, we need to productionalize this to a POC, like a Streamlit application, right, so it's like a Python app, web app, they can kind of do whatever they want, change the parameters, do it, it's self-service, but I didn't, I couldn't do that until I had like the foundation of like the model, so to speak, so it's like a ton of time went into that, that was not a good use of time in the sense of like tweaking every parameter for them, and so then I was like, okay, let's create a web application, and then that was a beast in and of itself, but then once that's up and running, you're kind of, you know, smooth sailing from there, so it's like 75 percent, and then the other 25 percent for me was creating a web application, or you know, some type of widget thing that the end unit, or the business unit could use to be able to do whatever they want, and I call it productionalizing it, because it's like, it's no longer in a notebook, but it's an actual app, right, and then another caveat is like the whole deploying, like when you're in the enterprise, you can't just, I would love to just deploy it in GitHub, right, and make it open to the world, but because it's proprietary information, you have to, that's another like a security thing, getting buy-in from your IT people, right, because that's a whole beast if you're in a big organization, you can't just like, you have to go through all these approvals, right, so that's another issue as well, and I can talk about it when I got Connect, right, because I'm not a Unix, a Linux admin, and so I had to kind of deploy it and learn all that stuff as well, right, and to be able to deploy it within, so I can just send a link to an internal Connect server, and then anyone in the world can access the web application easily, so if you don't have like the enterprise constraints, it's actually not nearly as bad as I'm describing it, right, so there's plenty of ways to deploy stuff freely within GitHub, but there's a huge amount of time building it, a little bit for creating the proof of concept, like the web application, and then you don't have to deploy it, but then after that, it's smooth sailing, because you just update it and push it, and that's why I kind of split it up like that.

Awesome, it reminds me of when people say like, if I had x amount of hours, I would spend 90% of that time understanding the problem, like you're doing that, you're understanding the problem, and then once you get that, and you invested all that, like building the thing is relatively easy. Yeah, yeah, yep, definitely, so a lot of it is understanding the problem, for sure, yeah, cool.

Convincing skeptical stakeholders

I have a question, how do you bring people, I had a client who was convinced that these bond metrics were just a few calculations in Excel, and it wasn't, it was a big deal, how do you bring people along who are really uninformed?

Yeah, so I think it is tricky, right, I think like sometimes you'll see, at least for me, you'll run analyses, and then it'll be like, almost like confirmation of like, oh yeah, I knew that all along, right, but like sometimes, at least for me, we'll have someone, I'll list that investor that, you know, the managing director hadn't knew about, but then it wasn't off the top of their head, right, and then for them, they're like, oh wow, that is interesting, and then that's relevant to us, but in terms of like different calculations, yeah, I think in your field, is there instances in which a value or some type of piece of information that surfaced that wouldn't necessarily be like, um, remembered off the top of their head, or is it, do you think it's just more of an understanding of like, what's the nuts and bolts of like, how you get to that answer?

I think it was, I think it's like, you know, calculate prices and yields and duration and what happens if you change the sensitivity to yield changes and prices, so it isn't stuff off the top of the head, maybe now that I've done it, it'll be more on top of my head, but he ended up stiffing me and not paying me at all, so I was really, really pissed, did an R Markdown, had the code for him, everything. Oh geez, that's a bummer, I'm sorry.

Yeah, well, so I want to get better, so it's like, the whole thing was convincing him that doing this work was more complicated than he thought. Oh, okay. That's where I need the skill, that's where I didn't do well.

Oh, no, I'm sorry about that. Yeah, sometimes, like, I think, like, in order to find that ability to impress the folks, is sometimes you kind of show, like, what is the normal route or the path in which they get their answers, and then you can kind of show yours, but in your case, it's a little bit tricky, right, or almost like, well, yeah, like, almost like a chart, right, it's like, here's how you normally do it, this is how long it takes, here's how my approach will take, it's more accurate, cheaper, better, faster, whatever, right, and then at least allows them to kind of see that. For me, what I, on my end of year things, I'll say, like, oh, I helped automate a process, right, that would take our analysts, you know, maybe 30 hours of work, and I just run, I run the script, and it just, it runs, right, and so then I can say, I'm saving the organization, let's say, 120 hours every year, that they can do other things, right, and through coordination, so that's how, at least, you almost kind of take the numerical approach and quantify of what you're doing to kind of sell yourself for, like, what you're doing to provide the customer

value, is what I found. And part of the story was to convince somebody that having reproducible research is a good idea, as opposed to, for a few calculations, so that's another thing to sell. Yeah, yeah, for sure, definitely. Thank you.

Yerim, you had a question about, kind of like a follow-up to Nancy's question, but also some thoughts to share here. Do you want to jump in? No worries if you prefer to pose in the chat, but I just wanted to let you know, Nancy, it looks like a lot of people are weighing in in the chat as well. Okay. I also wanted to add, just to let everybody know, if I miss anybody's questions in the chat, I don't have my Posit co-hosts here, co-hosts here today, so just feel free to raise your hand if there was anything I missed, so that I know. But thank you so much. I see, Darren, you just asked a question. Do you want to jump in?

Oh, I was not expecting that quick response. Oh, sorry, just saw it first.

Sean, I was curious, you had mentioned in your discussion, you know, your experience working with IT and having to do your own ETL pipelines, being the solo data scientist in an organization, and that can be a lot for just an individual person, especially someone kind of just starting out in data science. So I just kind of wanted to know more about your path and understanding who to talk to and how to get the resources you needed to be effective and, you know, do what you want to accomplish. Yeah, so are you asking more of like how I implemented Posit Connect and Workbench within the org? I mean, not necessarily. It's just, as a data scientist, there's so many things that, you know, an individual feel like they have to learn. And on top of that, there is this IT side of things that they may not be familiar with. And just understanding how data pipelines work or how data flows work, and that's a whole other area that, you know, some data scientists may not be exposed to.

Yeah, no, so truthfully, I think like I was kind of self-taught, right? And so I didn't know any better, right? So a good example is learning how, okay, so I have data that's like duplicated in Salesforce, like how can I make it so I can't dedupe it easily, right? And so then I created some scripts that would dedupe it on a nightly basis, but then to your point, like I need to be able to, let's say I have Connect or some type of process to automate it, right? Making sure you have buy-in from your IT folks. And sometimes with organizations, you are really restricted, right? To be able to even access the database. It's like sometimes you can't even get that, right? And at least for me, I was fortunate enough in that we were kind of building out our cloud infrastructure. And so I had a little bit of a hand in that, but I by no means knew like what was up. I think I talked about in my Connect talk for Posit that I was trying to update the server and I totally borked the image, right? And so luckily I backed it up. And so right now our production image is like the RStudio Connect backup. Like that's what we use in production now, right? Because I didn't know any better, right? And so it's just, I think it's kind of having that mentality to be able to move quickly and to kind of iterate and then know that you're going to make mistakes. But then I think it's highly dependent on your IT infrastructure, right? And then being communicative, right? So sometimes like their job is to make sure everything's secure and to not have data leakage or anything like that. And so making sure that you're on the same team, right? To be like, hey, look, I'm trying to, we're on the same team, we're trying to provide value to the organization and understand like what are their concerns, right? And then that you're not going rogue because what I've experienced is that that's not fun for anyone, right?

But in terms of, from my experience in IT, I think we had some consultants that helped us spin things up and then making sure that you had weekly or monthly meetings during the implementation phase so that everyone knew what was going on. And then as you work with different departments, you kind of earn that trust, right? And then trust, I think, at least in the enterprise that I've experienced, helps you get way further than trying to do it all on your own. But I think it is kind of highly dependent, but that's the experience that I've had that's worked out well. Definitely hiccups though. So thanks, John. Yep.

trust, I think, at least in the enterprise that I've experienced, helps you get way further than trying to do it all on your own.

Making an impact with smaller teams

Thank you. So I've had this come up a bit recently, but I think sometimes people see like stories about data science and huge companies or teams with tons of data scientists and think like that's so far from something that could be possible for my team if we have one person or three people. Some of my favorite stories are when I get to hear about like superpowers that Connect gives smaller teams, but this does not have to be a Connect focus at all. I'm just curious if you could share with us like some of the use cases where you think people could start making an impact right away if they're a smaller team.

Yeah. So I think for me, being able to host web applications, whether it be Streamlit, Shiny, or other like Flask applications securely, because that's like a big hurdle within, at least for me, is like I can't share it publicly, right? And so one of the quick wins for me is being able to host, whether it be a Quarto document, so or like our, I host, oh look at Apple, like doing all the little things. This is like the, look at that, see? I'm sorry. It enables it. So it allows you to be able to host things quickly, right? And so that people can access it in a secure environment.

And then you were asking, sorry about Connect. Yeah, no, I guess I was saying like not necessarily Connect specific, but just some of the like business use cases where people could start helping to make an impact right away. I think it's helpful sometimes, even if we're all working across different industries, to just get an idea of ways people are impacting the business. Yeah. Yeah. Yeah. So like a good example is someone creating basically like, I was helping one of our analysts out. I made a way to, like a pivot table essentially, but like where they just upload whatever CSP that they want and it automatically would do it for them. It was like a quick thing that I did for them that made their life so much easier, right? And it could, it would take me time to kind of do it. I was like, you know, I don't want to repeat myself over and over. So then I basically have been still to this day, it's like this little pivot table thing that will automatically do whatever they need to do, whether it be cropping photos for people, right? It's just like a little small toy example that does one thing and it's actually quite helpful, right? Because then it frees up my time, makes them a little bit easier.

It's one of those things that I think, it doesn't have to be complex because I think, you know, even I thought so. I wanted to have the most sophisticated app that has all the little, it's like a Ferrari, right? But you don't need that. You just need something simple, right? Just if it does one thing and it does it well and robustly, people will begin to use it over and over and over, right? And that's kind of, it's so funny. It's like, I was so proud of this app that I made, but it's like the simple thing, like the pivot tables, like people like that, right? And it's like, whatever, it's cool. It did not take me nearly as long, right? But sometimes it's those little small things that it does it well. And that's what, you know, can get people really excited about whatever you're producing.

Just if it does one thing and it does it well and robustly, people will begin to use it over and over and over, right?

I love that. How do you find out like what those things are that are taking people way longer than it should or like the opportunities to help somebody? Yeah. So I think it's just one of those things where it's like pattern recognition, where someone has come to you like more than three times to do this one thing because you do it well, it's very simple. Whether it be deduping or looking up or aggregating, like for us, we want to aggregate our quarterly, you know, metrics or KPIs. And so I do that a couple of times, like, you know what, maybe we should make this into a little self-service application so that they can kind of do whatever they want. They can kind of tweak things and then they'll be able to get out the analysis that they need, right? And so it's just kind of another David Robinson thing. Don't do it more than three times, right? Then create a function. But in this case, if you do something, maybe create a Shiny app or whatever app that can help them do whatever they need to do more efficiently.

Yeah. Cool. Thank you. Kevin, I see your comment here in the chat about having a repository of these simple apps. Would you mind just sharing a little bit more about what you'd like to see there? Are you thinking like examples of companies or? Yeah, no, I was just thinking that, you know, I think that's a great pivot table thing. You know, I think there are these things that non-power users, you know, end up having trouble doing. And across lots of industries, it's probably the same. There's probably, you know, a handful of tasks that people across all industries are doing in Excel that could be done, you know, with a nifty web app, you know, and yeah, I mean, it might be, yeah, it was just, I was just thinking it would be cool if an organization like Posit, you know, has access to all of us, you know, if somehow, yeah, I'm not sure how we'd prioritize them, because it wouldn't be helpful if you have a library of like a thousand web apps that do like all sorts of crazy things, you know, but if somehow, you know, it was curated, you know, really well, you know, then like, for example, this pivot table app that Sean's talking about, you know, we could probably all use that. And if it's like a Shiny app or something, everyone with a Posit server could just download it from GitHub and spin it up, you know, and try it. So yeah, I just, that was the idea. But again, it would take some curation, which would be tricky, but it could be valuable if it's done right.

Yeah, I love it. Yeah, I'm learning from all of you. And when we're talking about building business cases, I need to have these examples where now I can say, look, Kevin from the FDA asked about this. But seriously, if there are different apps like this or examples that you're you all are thinking of, you're like, oh, I have one I maybe could contribute, that would help me in building a case for something like this. So I started a thread in the chat if anybody wants to comment in there.

Working with IT

Okay, awesome. Let me jump to scrolling up some of the questions. So RM, I see you asked a question about working with IT. Do you want to jump in here and ask that? Sure, thanks. Just following up a thread that was mentioned earlier, what are you having tips or tricks or things that work for you and talking to IT? A common problem that I've seen is IT doesn't really talk data science and data science doesn't talk IT. Do you have to learn key concepts like what is a server? How do you secure it? Like basic stuff like that, that you may or may not know? Or conversely, do you have to teach IT here? This is what ETL, this is what it does. Here's why data is not clean. I'm curious to know what tips and tricks, you know, you might suggest or you think that might have worked for you.

Yeah, I think for me, understanding whatever protocol that they use, right, at least for us, remember, I'm not like a cloud security guy. Learning that we use Okta, right, and making sure that all the things that we had were secured via SSO, right, when we were spinning up Posit, Connect within our organization, making sure that we had Okta authenticate everything and making sure that using like the key words or buzzwords that they identify with, because to your point, I don't necessarily speak IT natively, right, although like the security and then conversely, they don't speak data science necessarily, right, but to make sure it's like, hey, I understand, this is a high concern for you. And I share that as well. And here's what we've done to address that. Does this sound reasonable to you? Right. And so like, almost meeting them where they are is quite helpful, even though you may not speak exactly the same language, right. So like my understanding with Okta and all that is like, you know, we're using that as the gatekeeper to authenticate everything. And once you you use that and you're compliant, then IT people were, were relieved, right? Or like, okay, this is good. But then I think your mileage may vary depending on the organization. I don't know how it's like, like the FDA, right? I don't know, things like that. In terms of like, how it is, right. So it's, it is context dependent. But in terms of the general tips that I found that helped me out, it was like, communicating, meeting them. You know, even if you know, it's like, it's secure, like making sure that they understand that you're not trying to go around them, right. To do that. Right. Thank you. That makes sense.

Tackling data silos

Thank you. I see a few anonymous questions I want to go over to as well. And one was, it says, Thanks a lot. Can you share how you tackled data silos and access company data sources? Did you handle it alone or seek support from other teams?

Yeah, so data silos can exist. I think it that's, um, for me, it's like, it was not necessarily data silos, because all of our our stuff is in the cloud. So we use Google Cloud for Google Cloud Platform for our stuff. But for me, it was data was siloed in the brains of like the business unit. And so what we've done is do informational interviews where you kind of, you, you, you go tag along with them, right, to kind of see like, how do they go about doing

their job, right. And so by going to, let's say, the business development team, or the investment team, and kind of shadowing them, you understand their either their pain points, or like what they can, what you can do as a data scientist to kind of help make their job easier. Right. And so that's kind of what what I had to do to understand that, oh, they're really looking for this information. And it is there, it's just maybe not easily accessible to them. Right. And so that's how you do that.

Okay, but at least for me, all the data is in like BigQuery and all the because we feed everything into the data warehouse. And then we do all the transformations and send it back to like a clean layer. Right. But in order for me to know what to extract, because there's so much raw data, right? You almost have to understand what is the business trying to achieve? Right? Because sometimes I've done things all the time where I do stuff. That's what I think is like addresses what they need, but it's not exactly what they will want. Right. It's like they may ask for ABC, but they really need DF, right. But you don't realize that until you kind of talk to them, even though they say, I want ABC, they really need DF. And you don't get that until you actually kind of understand things.

Because like, we are working with everything. But that's what I've noticed is like you deliver what they asked for, but not necessarily what they need, if that makes sense. Right. So it's like, you only get that by kind of iterating understanding where data may be or whatever problem that they're trying to solve in terms of silo. And then you'll be able to integrate it that way. But I realized that that may not be the case, because some enterprises have true data silos where it's like a totally different server. One is on-prem, one is in the cloud, one you don't have access rights to if it's like healthcare data. So I can't speak to that, but I can at least speak to my experience.

Iterating quickly to understand what stakeholders really need

How do you figure out more quickly what they actually need if it's not what they told you they want? Yeah. So I, you know, because I, you know, I was in grad school, I was, I'm used to working late night hours and just working on it for just brute force figuring it out. But what I had to realize when I pivoted the industry is being able to iterate quickly to be able to show them like a janky looking report with some number. It's like, it's like, is this right? Is this not right? Is this like in the ballpark? And then getting feedback or directionality because, you know, I do like that, like the shiny well-polished application, but it's like, it'll get to that thing where you work on it, but then it's like, they're like, this is not what I wanted, right? And so you're better, you're much better off showing something that's really rough, but it's somewhat tangentially relevant to them. And then iterating again, then to like take all the time, make it super beautiful looking, but it's like not the values of the information that they will want, right? So that was kind of one thing I had to kind of unlearn when I was in graduate school to the industry is like to, to make sure you're getting feedback from them.

Being able to iterate quickly to be able to show them like a janky looking report with some number. It's like, it's like, is this right? Is this not right? Is this like in the ballpark? And so you're better, you're much better off showing something that's really rough, but it's somewhat tangentially relevant to them.

Yeah. That's great. Thank you. Alan, I think you have a question in the chat. Also a comment. So I'll just have you jump in. I wasn't reading it all yet. Oh no, I don't want to be the, my question is more of a comment guy. No, no, not at all. I just can't use both parts of my brain.

I think I have two competing things going on in my brain. One of which is I love the idea of the like pivot table use case and all it opens up of making nimble, flexible ad hoc tools that really meet a business solution. And I love that because it's elegant and it's an exciting way to work. And at the same time, I think it competes with pushback about like proliferation of solutions and a different thing, a different avenue of doing work, which would be like saying, Hey, the analyst has just got to get better at RBI or, or, or whatever. Right. And so I'm really curious about what are the stories to tell that motivate the kind of work that we all really, really super love versus that other, that other story. And so I'm just, I'm curious if Sean, if you, or if others have that success story that motivates that way to work, as opposed to the, like, well, no, you use the enterprise BI app channel way to work.

Cause I think those are in somewhat tension. Yeah. I totally agree. I think it was for us because we're a little bit smaller, right? We don't have that problem yet, but I can definitely envision this being something that is, if you have like a larger team where, yeah, to your point, the analysts or whoever should get more proficient. But in my, at least for my instances, it's, it's our admin that needs, that just wants a quick calculation to get, to get on, to do whatever they need to do.

But I think it's, that's one of those things where, at least for me, if it's in connect, you can kind of look at the usage metrics and theoretically look at the logs, right? To kind of see like, Ooh, maybe this person. And I think you have a conversation, right? It's like, okay, it seems like you, this is really valuable to you. And then you can kind of address, like, maybe they need to get better. I think you ought to be strategic about that. You don't want to be like logging people's activity. But I think one, another instance is let's say you have a suite of really small toy, shiny apps or whatever for pivot tables and all that you can almost making that like a featured application that allows them to do whatever they need to do. So let's say they need to pivot, they need a crop and they need to, I don't know, make sure the data is not missing, right? You can almost package that into a larger application that allows them to do that. But I do, I do see your point about like, there is a time and a place to use the enterprise preferred business intelligence tools, right? And that's tricky, for sure.

Developing intuition as a data scientist

Anonymous question was, can you share how you developed the intuition as a data scientist or statistician to make certain decisions? Intuition? Yeah, I think for you guys, for me, you get like that code sensor like that, you get the inkling for me, what the business would want, right? And then from there, it's kind of testing the water. So what I do is I have folks that aren't necessarily super senior managers, but I have the junior level people that are in the, you know, working on the business, and I will send them little snippets of like projects, works in progress, right? And then I almost use them as a barometer, like, am I in the right direction? And then if it is, then we kind of go there. So that's like, when I did that Suji wrapped, right, I went to a couple of our associates, and I was like, hey, what do you guys think of this? Does this look and feel right? Because sometimes when you're presenting to management, if the values are off, they're gonna like, why is this off? Why is it, right? And so you kind of, you go to the people that are a little bit more junior, and you ask them, like, hey, does this feel right? Does this seem right? Does it not? And what would you want on, like, the next slide or whatever, right? And so then you kind of use them to be able to guide yourself on producing different applications or products, what data products.

Managing long-term vision with management

Thank you. Kevin, I realize I missed a question from you earlier about getting feedback from stakeholders. Do you want to add anything? Yeah, sure. Yeah, I think, I think it's the question, my question really what I want to ask is, like, how do you manage like the kind of long term vision, right? When when maybe the management doesn't, you know, really know what's possible. And they Yeah, I find, I find myself in this position sometimes where I'm, I'm trying to figure out how to, like, where long term, we want to prioritize and build things, but the management doesn't really know what's possible. And so it's hard to get from them, like, what would they prioritize? You know, like, yeah, how do you kind of get priorities for management that's, that's not in touch with, with what's possible, if that makes sense?

No, for sure. I think, you know, what I've done, right, is I do whatever we is asked of me, right? So let's say it's like the XYZ KPIs, you do that. But then sometimes I'll work on a side project. And at least for me, I was doing like a relational graph network this year for a wrap. And I kind of slipped that in as like a little things like, I think this is pretty cool, they're gonna think it's relevant. And you do everything, you're not going rogue, per se. But you're just adding that little, oh, look at this cool thing, right? And then you kind of you try to find the most, not shiny, but like more like interesting thing, if you can, and insert it. And then you that will pique their interest, right? And then they'll ask you for more. And then you almost kind of get that permission to, oh, that was just something I spun up, you know, last afternoon, right? And then they're like, oh, yeah, let's do that. Right. So it's kind of like the indirect way of doing and I'm not advocating for like, disobeying and your manager or anything like that. But like, what I'm saying is, you know, they may not know what's possible, right? And so what we're trying to do is show them like, hey, this might be possible. Would you like me to spend more time on this, right? And you're still do whatever you need to do, right? That's the beauty of having code is you're able to use something and recycle it and improve it. But then at the same time, you can kind of work on other things and maybe slip that in and then get them to to buy in.

And so that's like truthfully, that's that's basically the journey that I've been been been going on is like, I knew what I wanted to do, but like there's certain things that need to be in place in order to do that. Right. So if I wanted to do a graph network, I need to make sure the data is correct. Right. Yeah. Get some way to surface it to them and then do that. And so then you do those things and then you kind of slip it in once all those prerequisites are met.

But no, and my question is less about, yeah, getting giving yourself permission to do that, because I totally get that. Yeah, you got to do it and show them. Right. But it's more like just a really big picture. Like, how do you how do you get input on the vision casting for that when you know, because because obviously you can make your own vision of what you think this graph thing, you have a whole vision for it and you can go forward with that. And that's awesome. But ideally, when you're working for an organization, you do want to somehow get input or some direction from, you know, from the leadership and not just always be steering them.

Yeah. Yeah. I think it's one of those things where you try to have that open conversation with the management and where you're getting like, oh, what do you think about next year, us creating like an S2G chatbot or we're using like a link chain LLM type of stuff. And, you know, they may not necessarily understand, be like, oh, a chat GPT for the organization. Right. And so then they're like, oh, yes, we would love that. Right. And it's almost kind of like, you know, it depends on your relationship with the management as well. Like, yeah, no, that's really helpful. Having a conversation would be like, you know, let's let's just brainstorm what we call like, you know, at least my relationship with the management is like that where you can kind of like, hey, what do you think about like a chat GPT for the work? Right. Like it's going to be a huge effort. Like, don't get me wrong, but you're having a conversation. It's like, is this something worth pursuing? It's just not.

And then from there, you kind of understand their priorities as well. But I acknowledge that it's that may not be the case for some people. I would imagine the FDA, you can't just waltz up to the administrator. Right. But, you know, I'm trying to understand and meet with the decision makers and kind of ask them about the vision and slip some, you know, would something like this be useful or not is like what I was thinking. I don't know. Yeah, I think the brainstorming approach makes sense. Yeah. Let's get. Thanks. Love it.

Closing remarks

Well, thank you all so much for joining us today and for all the hangouts that you've joined us in in twenty twenty three. Thank you so much, Sean, for sharing your experience with us today as well. Thank you for having me.

Wishing everybody a great holiday and we'll see you all back in the new year. So we'll be back on January 11th with Marin from TD Bank. And I've been listening to a lot of podcasts lately where they do their credits at the end and I realize I don't do that here. So I do want to say thank you as well to all the people from Posit who make the data science hangout happen. You may notice I occasionally have a different co-host on helping me. So thank you to Hannah, Tyler, Robert, Curtis and Catherine, who will jump in to do that, as well as our creative team who helps us make sure we get these recordings up and the upcoming speaker list. So thank you to Olivia, Maya, Green and Margaret. But have a great holiday season and a happy new year, everybody.

Thank you. Thank you. Thank you. I'll.