Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Kevin Wee’s Tableau Journey in Visualization and Innovation
0 Playsin 2 days
Happy New Year, everyone! I hope you’re having a great start to 2025 and are poised to have a happy, healthy, and successful year. I’m excited to start off the podcast this year with my friend Kevin Wee, a Tableau creator and visualizer, who shares his journey into data visualization. Kevin’s Tableau journey starts in 2019 at Purdue University where he faced challenges working with data, but grew in his ability to create interesting and engaging dashboards through creative projects and community involvement. Kevin and I talk about his approach to Tableau, how he incorporates other tools like Figma, PowerPoint, and Excel, and how he is using AI tools to help create better and better data tools. I hope you’ll enjoy this week’s episode of the show!
Keywords: data, data visualization, PolicyVizPodcast, JonSchwabisch, KevinWee, Tableau, DataVisualization, TableauCreator, PurdueUniversity, BiomedicalProject, DataCareer, TableauPublic, MakeoverMonday, TableauAmbassador, DataTools, PowerPoint, Excel, TableauCommunity, Figma, InteractiveDashboards, TalkDataPodcastToMe, DataPodcasts, ChatGPT, SalesforceIntegration, AIDevelopment, DataIntegrity, DataHarmonization, DataAnalysts, FutureAmbitions
Subscribe to the PolicyViz Podcast wherever you get your podcasts.
Become a patron of the PolicyViz Podcast for as little as a buck a month
Follow Kevin’s Tableau Public page and check out his website.
Follow me on Instagram, LinkedIn, Substack, Twitter, Website, YouTube
Email: jon@policyviz.com
Recommended
Transcript
Welcome and Podcast Introduction
00:00:13
Speaker
Welcome back to the policy of this podcast. I'm your host, John Schwabish. Happy New Year, everybody. I hope you had a nice, safe and healthy New Year and holiday season. I hope you got a little time to rest and relax and gear up for another great season of the podcast. I've got a great set of guests for you coming up over the next few months. We're gonna talk about different tools to create your data visualizations. We've got folks coming on the show to talk about how to give good presentations, how to create better data visualizations, how to work with qualitative data and lots, lots more.
Introducing Kevin Wee and His Tableau Journey
00:00:48
Speaker
And to kick off 2025,
00:00:51
Speaker
I'm excited to welcome Kevin Wee to the show. Kevin is a tableau and data visualization creator. Kevin and I have known each other for a while now and I finally was inspired to invite him to come on the show because he created this really cool dashboard a couple of months ago.
00:01:08
Speaker
on data podcasts of which, of course, this podcast was featured. So I'm very grateful to him for including me. And so I invited him to come on the show to talk about his Tableau journey, how he thinks about using Tableau, how he creates his really fabulous backgrounds and foreground images in the tool. And now that we're at the beginning of the year, what he thinks the tool is going to do and improve or not over the next year and beyond. So, no further ado, here's my conversation with Kevin Wie on the first episode of the Policy Viz Podcast in 2025. I hope you enjoy it.
00:01:49
Speaker
Hey Kevin, happy holidays. Happy new year. How are you? Good to see you again. It's been a few months. It's been more than a few months. I think the last time you and I met was in the outlier conference. Yeah. Well, that was probably first. and i Then we saw each other at Tableau. And then before that it was GRC and then at Tableau conference. That's right. That's right. GRC. I forgot about that. Yeah, that's right.
00:02:13
Speaker
All right. Well, great to see you again. I hope all is well. So we're going to talk about Tableau today for folks who always want to hear about your cool work. I thought we would start with your Tableau journey because everybody has like this unique, interesting story about how they got into Tableau and why they got all excited about it. So I want
Advice on Building a Data Visualization Portfolio
00:02:30
Speaker
to hear about that. And then we can talk about some of the stuff that you've been working on.
00:02:32
Speaker
Yeah, sure. um Regarding my Tableau journey, I started ah learning about Tableau in 2019. I was a graduate student at Purdue University in the Department of Chemistry. However, my research was on scientific visualization. Effectively, I was evaluating a specific biomedical animation produced by Harvard Medical School, University of Toronto, and UC Davis. They co-produced.
00:03:02
Speaker
an animation, um and I was the evaluator of the animation. um By the end of ah the data gathering process, I got a lot of data about how the animation was made, how students perceived animations, how teachers wanted students to learn from the animation, got all this data, had no idea how to analyze it at all. like at all So I was exploring different analytics tools, quantitative and qualitative options. Eventually I bumped into Tableau. um um It was recommended by one of the data visualization professor that that I met in an academic conference. And then I took her class in 2019. And then after that, just started to use it to build some charts for my own project, and then the pandemic hit.
00:03:54
Speaker
um in 2020. And then there was also the time where I realized that I really need to find um my next career. I knew that very early on that I didn't want to work in academia. um So in the pandemic, when everybody, when Tableau Conference was online that that year, ah that's where I came across the term Tableau Public.
00:04:20
Speaker
Makeover Monday. ah Everybody recommended me in order to find a data job out of academia, it would be best if you could have a portfolio. So I started doing Makeover Monday um and then gradually realizing that I was actually quite good at that thing. And then um By the beginning of 2021, I was reached out by someone in the data visualization society, ah specifically in a Chicago location tab. Because I've been sharing my personal portfolio work in the Slack channel. So he reached out to me and said that, oh, by the way, I was also from Purdue Chemistry. And you were also from Purdue Chemistry, but you're not doing chemistry stuff on the internet. What's up?
00:05:10
Speaker
And then we talked long story short after a long two and three hours chat. He said that, by the way, I'm currently working the credit card company and we may have an opening position if you would like to test the water with us. That's how I got my job and I just passed my third year anniversary recently. So um so that's my career path. So I switched out of academia.
00:05:33
Speaker
But at the same time, I still do a lot of things with the Tableau user community, or ah some people that are familiar with the term, it would be called the DataFam. So I was pretty engaged with the DataFam. I started giving talks in different Tableau user groups, or a Tux for short.
00:05:52
Speaker
and um suddenly everybody want ah not everybody that's very you know not humble of me suddenly i received a lot of opportunity uh to give talks in different occasions and that's um i think in 2022 i was ah chosen as a tableau ambassador and then starting this year i was chosen as a visionary based on all these works that i've been doing so Yeah, that's a very long answer, too. No, that's great. Well, congrats congrats on all that. I wanted to ask about the portfolio building, because this is a question that comes up a lot for people who are starting out or maybe they're shifting gears in their career. And you know one of the questions I hear the same kind of thing is like, you know how do I kind of get started in in the field? And so how did you think about building your
00:06:47
Speaker
DataViz portfolio? was it were you Were you sticking with a single tool? Were you just like working in the Makeover Monday and just trying things? like where And where were you posting things? like so So for folks who are kind of in a similar boat as you were, what would your recommendations be to help them build out like their DataViz portfolio?
00:07:11
Speaker
I think um first of all, I think a lot of people when they started out, they were overwhelmed by all the options and platforms out there. It's understandable because especially nowadays, different tools, ah a lot of tools are very easy to use and a lot of tools are free. So how do you choose? I think my answer is it doesn't matter. Stick with one and be good at one to begin with and then branch out.
00:07:38
Speaker
um Of course, to me, a lot of people didn't know that even though now I was known for a Tableau, before 2019 at Purdue, I was known by the person that can do everything on PowerPoint. So I was on PowerPoint a lot. I was on Excel a lot. It's just that I realized that PowerPoint and Excel is hard to manipulate a large body of dataset, especially non-numerical datasets. um That's why I branch out to different options.
00:08:09
Speaker
and um And that's why i I've learned other softwares. um I was actually actually also exploring a rawgraph.io for a couple of months. ah But then I realized that it's easy to start with something, but it's hard to customize um the chart compared to using Tableau. But okay, go back. um I started using Tableau because number one, it's free for students.
00:08:34
Speaker
So that's easy. That's free for students. And that's also what I've heard. Very famous on the street. By that, I mean ah in the industry. So um Purdue gave annual all free license to students, so I leveraged it.
00:08:50
Speaker
and Also, I also learned that Tableau Public, it's a, you know, freely used platform website for you to put your work up there and then share your take on different data sets. So um that's how I get started. And also, a lot of people also ask, like, there's so many different, you know, challenges for data.
00:09:11
Speaker
out there. And i my recommendation is always, um if you just started, just pick one. And I would say maybe one of these two, Makeover Monday or Back to Vis Basics, because these two are just enough for you to warm up. And once you get into the cadence of creating good quality charts with these two challenges,
00:09:35
Speaker
you can branch out. You will know when you're ready to branch out. Don't try to do everything to begin with. Stick with one thing and do it until you can repeat it without too much of an effort and then you branch
Creating Consistent Dashboards with Figma and PowerPoint
00:09:47
Speaker
out. And in Tableau, I think unlike many other tools, um I think one of the one of the, you already mentioned the data fan, but I think that's one of the reasons why people you know really enjoy working with the tools, that there is this community and you have this opportunity to sort of build your profile on Tableau's public and so on the page and so you can sort of build out your your portfolio. Is that where you pointed people or did you set up your own separate website to point people to? Again, I'm thinking about someone who, you know, wants to break in, they want to do dataviz, say, you know, they're they're applying for jobs, they want to do dataviz. Like, should they point people to
00:10:27
Speaker
their Tableau public profile and say, I, you know, I'm really good at Tableau. This is what I do. Here's my profile. Or do you think it makes more sense to have sort of a separate place where people can go? Let me tell you what I did. And then we made talk about my recommendation. Yeah. I, I started getting my own Kevin, we dot.com in 2018, but I didn't actually use it much. I was just more like, I don't want anyone else to have it. so that just i Yeah.
00:10:53
Speaker
yeah And then after that, i if you go to my KevinWe.com and scroll to the bottom, you will find that at the first couple of blocks that I've made were my own makeover for some charts with the data that I gathered using PowerPoint.
00:11:09
Speaker
e And then I not just show before and after, I also explain um why I made a certain design choices. It's a good starter, but I am a very slow writer. I realized that ah very, very soon that I, if you ask me to write something, you're going to wait for at least a month. So I noticed that about me. and Meanwhile, make over Monday kind of force you into a weekly cadence, right? And then you have to keep on producing the charts and then only justify it when someone asks of it.
00:11:37
Speaker
so So that became some something that I am comfy with. um right If you're a nonwriter like me, if you like to show more than you like to tell, then maybe Tableau Public would be something that you could direct people to. But let's say if you're someone like, you know,
00:11:56
Speaker
And you can write a lot about your justifications. You can cite different articles. Then ah maybe a blog would be something that is more suitable to you. Or you like to sketch. You like to show the sketch and then the viz. Then the blog would be perfect for you. Yeah, that's smart. Yeah, sort of the ah talking more about the the process or in addition to the final thing, talking about the process. Yeah, that' that's really interesting.
00:12:22
Speaker
um so So one of the places where I, in my own Tableau work, found your work super helpful was in these, I don't know what you call them, but sort of foreground or background images that you can put on your dashboard, you know, sort of like a skin. um And I was hoping you could talk a little bit about how you think of developing those and then how you actually build them. I know you use, I think you use Figma a lot. The last one I did that needed I needed your help on, I was doing in PowerPoint, but maybe you could talk a little bit about how you how you think about these images in Tableau and then how you go about building them out. Yeah, well, first of all, um it's I think it's a gap that we see in the existing Tableau.
00:13:09
Speaker
ecosystem that it doesn't have too many graphic manipulation interface. yeah As a result, even in 2020, I already started seeing people using PowerPoint to create images and then lay it as a background or a semi-transparent foreground of a dashboard.
00:13:32
Speaker
um With that in mind, and also if you go to Tableau Public, you will see a lot of times people's visualizations, they're clearly not out of the box functionality of Tableau. Be it title, font, or like sometimes you see some very dramatic shading or gradients on the infographic. You know that these are not built just using Tableau. So, and gradually I've heard that the term Figma pop up and then And then I've heard that it is one of the commonly used software for people trying to implement graphic design and tableau charts. um I think I learned it in early 2021. So it's still rather rather short period of time.
00:14:23
Speaker
and um I just try to play with it. At first, I just try to use it to create, you know, rounded rectangles to place my charts, just make it look a little bit more modern. And by modern, I mean, you know, more than four years ago. I don't feel like that they're modern anymore. But yeah, it's just to to do something to stand out a little bit without distracting too much. But then later, when I joined Discover, the the company that I'm currently working with, I got several big projects where I got to build dashboards from scratch. So I did tell my manager that um
00:15:01
Speaker
Can you let me try to use Figma just for one project? Let's just try to do it and I'm going to use it on the background and maybe lay a semi-transparent annotation layer in front of the foreground and the users can click show and hide of that annotation layer at will. Let's just sandwich this and see how it goes.
00:15:22
Speaker
Business partner love it. I think it up to the SVPs. Everybody was blown away by that. and then And then I realized that if I set up this background image and semi-transparent layer well, I can quickly create different iterations of them.
00:15:38
Speaker
Therefore, I started to be able to mass produce a lot of different dashboard layouts in a very short period of time. And they're all visually consistent with good organizations and layouts, and everything's perfectly aligned. Right. Everything's perfectly aligned. It looks like it's coming out of a united team instead of everybody's doing their own stuff. So after doing that, I remember within my first five projects, I was able to create at least 20-ish dashboard pages, all with very consistent view. And then everybody's were blown away by that. My boss told me that, I thought you're going to take three months for this. You finish everything in a week. I was like, I didn't know that I was supposed to do that. Yeah, I had all this time. Yeah. So did you, but to set that up,
00:16:26
Speaker
You need to know your dashboard size and layout before you go and build the skin, right? So yeah, like what is your creative process when you're doing that? Are you like, I'm just going to build what I think is the best dashboard and I'm not going to worry about the aesthetic. I'm not going to overly worry myself about the aesthetics until like the very end when I can go build in these backgrounds or foregrounds.
00:16:50
Speaker
Yeah, so as ah two part of the question is, first of all, let let me answer the dimension first. yeah So I test the dimensions of it. First of all, we're using the fixed dimension. So we're not doing flexible. um The thing about flexible dimension is that everybody has a different laptop and everybody's at home. They have different screens. If you let them go whatever they want, nothing's going to align. It's going to be a disaster.
Importance of Templates for Visual Consistency
00:17:14
Speaker
What I notice is as long as the width and the height is in the golden ratio, it would fit most of the screen in the good degree. e So that's good enough. And also um after several trials, I believe we set the width to be 13.8 pixel.
00:17:34
Speaker
The reason why it's 13.8, not 1300. Yeah. It's because 13.8 is divisible by 12. So in the middle, you can put one, two, three, four, six charts, and everything was still aligned with without, um, distribute evenly without, um, the need of decimals. e Yeah. Right. And also it's making it very easy to see whether someone stole our design because there's no way you come out with 1808. Yeah. No, 1308, I'm sorry, 1308. So that's the dimension problem. Second, aesthetic part. I think a lot of people thought the reason why we use Figma or any graphical software to create a background and foreground was because of the aesthetic. Don't get me wrong, I do think they look good. yeah But to me, they are first to have a large team have a consistent visual language.
00:18:31
Speaker
And now I already set the template for you. Now all you have to do is just reorganize the cards for the charts. You don't have to worry about what's the color for the font. You don't have to worry about the color for the site navigation bar. You just need to care about the chart parts. It kind of reduced the workload for our day-to-day analyst. So that is to me more powerful than aesthetic. Aesthetic is just that I don't want to create anything ugly, but that's not the sole reason. Yeah, no, that that that makes total sense. I mean, I know lots of people have like mixed feelings about templates, but i think um I think a lot of people who have mixed feelings about templates haven't worked in large organizations where people need to get stuff done to get out the door. And yeah like you said, you you lower the barrier for folks to focus on the data work and worry less about the branding and the and the consistency.
00:19:29
Speaker
So I reached out to you to do this show.
Kevin's 'Talk Data Podcast to Me' Project
00:19:31
Speaker
We've talked about this. I've talked about having you on the show in the past and never sort of got around to asking. But then a couple of months ago, you published a big dashboard called Talk Data Podcast to me, um which I kind of think went like as viral as things can go in our database community. So you know that's that's pretty cool. um And so I thought maybe you could kind of walk us through that that project.
00:19:55
Speaker
Yeah, it's, um well, first of all, it was ah my personal like passion project. No one's, you know, convinced me to do it. um I listen to a lot of data podcasts while I'm working out cooking and driving. It's just a perfect time killer for me.
00:20:12
Speaker
um But then I realized a lot of people, especially people that just, ah not just people that just started practicing in data field, a lot of people were, even those that work in COE of a company, um, they didn't realize that, Oh, there's something they call data podcasts. e And because to them, data visualizations or statistics, is something that's you have to do it on the screen, right? So what is it to talk about? Right.
00:20:39
Speaker
so um So I always wanted to compile a list, maybe on my blog, to show people these are the podcasts that I listened to, these are the graphic design YouTube channels that I followed. That may be another story. But I want to show people this list and then try to help ah people to get into the community or or the network more. But then it would be very boring if I just give you an Excel spreadsheet with you know hyperlinks. So meanwhile, i was trying to I've been hearing people talking about, oh, you can use chat GPT to do coding a lot. So I thought, let's just try using it. So there's one weekend I talked to my chat GPT. And I was like, oh.
00:21:27
Speaker
write me a web scraping Python codes ah to get every episode for Data Plus Love, ah the podcast channel. And then after that, I was like, okay, I like the data set. Now try to rewrite it to export it in CSV. Now rewrite it to get API credentials from my Spotify channel.
00:21:49
Speaker
um now rewrite it for 10 different podcast channels. And then step by steps, I got a huge dataset. yeah And after that, I just explored it.
00:22:02
Speaker
and then put it to Tableau. And then, of course, use my Figma magic to come up with a layout, the design, the color palette, and then ah put together the dashboard that you saw. ah Talk Data podcast to me is actually you know coming from the euphemisms talk data to me. and And that's before that was talk dirty to me. Yeah, talk data to me. And I was like, OK, how about Talk Data podcast?
00:22:27
Speaker
to me. And yeah, and um because this is a personal project, it's not a business project, I try to make it a little bit fun, right? ah First of all, dark mode, not line mode. And also, if this is a business dashboard, everything will be rectangle, right? It will be rectangle bars and everything, but it's I want to make it fun because it's for myself. So I make it, you know, look a little bit like the synthesizer. Do you know the synthesizers for DJing? Yeah. Trying to make it look like that. And then also leveraging the multi-hue sequential color palette, Verdist. And Verdist happens to have purple and green, which is the which are the color for Spotify and Apple Podcasts. So it's like, okay, everything's aligned. Let's do it. Yeah.
00:23:11
Speaker
Yeah, so yeah, that's the final outcome. that so did you Did you discover anything about your personal podcast listening habits by by actually looking at the data? So no, I did not look into which podcast I've played. um Yeah.
00:23:28
Speaker
for several reasons. Number one is I noticed, because sometimes I do it on Spotify, sometimes I do it on Apple Podcasts. So it's not gonna be complete. And second, I don't i don't know how to solve this yet, but sometimes when you play it, it would replay.
00:23:44
Speaker
And if you did not replay it through, it would say that you didn't play it through. I already listened to this thing more than once. So I haven't solved that yet. So, uh, no, I did it. So for that project, I didn't include any of my personal listening history in it. But I do know that like some of the famous one, I did a plus love, uh, policy of this podcast. I have to be frank. I didn't listen to all of them. I only pick the one that more on the graphical side. Yeah. Yeah. Yeah.
00:24:13
Speaker
Well, not just more on the graphical side. You also have a couple of episodes with, um, you know, Jen Christensen or you also talk about qualitative data. Those are the ones that the specific topics that I like. So I rhyme with them. Yeah. And yeah, so these, that's the final outcome that you see. So I saw it like everywhere. I saw it on all my channels. And so what, what was the feedback? Like what were you hearing from people when they, when they saw it and they started playing around with it? I, I'm trying to think.
00:24:44
Speaker
I think number one is a lot of people appreciate that I am like, they never heard of a lot of these channels before. So it's like in terms of knowledge, they right appreciated that now we have it. And I think other than that, also those that are very, you know, good in data visualization, they would appreciate the UI interface, like how slick everything was, especially when you click the bars around the dots would just move very smoothly. they' They're more, you know, technically impressed. um Uh, but also I also have a couple of people realizing that oh a lot of the channels has been discontinued Right. So they're like, oh, I have a lot of episodes But now you're showing me only have like a couple couple was that because I only picked the last five years because that's you know When hell breaks loose in the whole world. That's when I started to listen to podcasts so um, so yeah, it's I got a mix uh back of um reviews but
00:25:36
Speaker
And also also, a lot of people were impressed by the color palette. Apparently there's still a lot of people didn't realize multi-hue sequential color that are color blind friendly. huh It's a thing. So maybe maybe you could invite people to talk about this because i think sure I think we need to have more conversations on other than the, you know, Inferna, Magma,
00:25:57
Speaker
Um, plasma and various, uh, what else that we can offer to the world. Yeah. Yeah, for sure. Um, okay. So I want to ask you a couple of specific things about, about
Tableau's Community and Future Prospects
00:26:08
Speaker
Tableau. So I want to start with we're in the new year here. Um, what, let's start with what do you like the most and what do you hate the most about, about Tableau?
00:26:17
Speaker
I see, I see. Let's start with what I like most about Tableau. I think what I like most about Tableau is that it has a very mature and well built network of users that are willing to help each other.
00:26:32
Speaker
I've i tried to check this with, you never heard this about Excel. No, right, that's right. You never heard this about R or Power BI is trying to do something similar, which I appreciate because I think yeah in order for the Tableau users to get stronger, you need to have someone that's something else that's equally stronger to to to feed each other is the energy, right? Yeah. So that is something that I just find it amazing. I am part of this, you know,
00:26:59
Speaker
community, of course. And, and I think just the fact that every time you throw a questions out there, someone would immediately jump into helping you without belittling you. um Most of the time, it's a, it's, it's a good thing. That's also why I always feel very welcome when I step out of academia, people not just tell me that, Oh, this guy's been living in ivory tower for his whole life. What does he want to do with us? I've never.
00:27:26
Speaker
receive those vibes in the Tableau user community. So that's something that I like the most. um What I think may be something that's like, can we improve the most? Maybe I think currently there are a lot of discussions on, well, first of all, I think there has a lot of opportunities to improve between different Tableau softwares, like Tableau desktop, Tableau public, Tableau server, Tableau cloud. If you use any two of these four combos, you will realize that a lot of things that can only be done in one of them. and It's very annoying to have to figure out what they are.
00:28:06
Speaker
Yeah, I think that's I think that that's a really good point. But I think that's also true across lots of tools. Like I was like, I was just working just finishing a paper with colleagues at work. And we were trying to do things in word online, like on the equation editor. And I know there are people rolling their eyes at doing equations in Microsoft Word. But like, You could not do it in the online version. So I do it on the desktop version, but then when uploading it back to the online version, it would break. And so I think that kind of feels like a common problem right now. Yeah, it's a common problem. However, does.
00:28:44
Speaker
does the user base mainly focus on mainly focus on one of them? right like if ah Let me use Figma as an example. Figma has a desktop app and it also have a web app, but I bet like most people are using the web app. yeah So that one better have the most functionality compared to the rest. I feel like because Tableau's user base, like day-to-day users, not those that use for fun, um they are quite evenly across all these channels, all these different applications. So how do they balance that? right And if they were to always highlight these four products, then they need to do they need to improve the consistency between them. And I think there are a lot of, I know that there are a lot of efforts at the backend to try to, you know,
00:29:35
Speaker
make these consistent. I just wish this is just me wishing it to come sooner. How about that? Yeah. Yeah. Yeah. I hear that. I hear that for sure. Yeah. um So at the beginning of the year here, where do you think Tableau is heading as a tool now is, I mean, we could call it a company as part of Salesforce. Like where do you think it's heading and where would you like it to head? I mean, obviously having these, as you just mentioned, having all these different versions, a little bit more, uh, better handshakes between them, but like, where do you see it going? Where do you hope it's going to go? Do you have any fears about where it's going to go in this new world of, you know, AI cloud within Salesforce? Like what do you, what are you thinking is going to happen over the next, next year?
00:30:22
Speaker
i think I think inevitably it would be very, very AI-focused for good reasons, right? Because it's a new tool. Everybody wants to use it. um However, I think what I would like to see more in the future is that with the AI, would you be able to develop a system telling the analysts how to place themselves in this using like a lot of Tableau's recent demonstration, you will see that it's all about executive using the AI ah in Tableau to figure out what it is. But we know that in real life data sets often very inconsistent, very dirty, ah very problematic. So how
00:31:12
Speaker
can analysts join this conversation and become a safeguard um to making sure that the data that they receive or interprets are proper. I think that that conversation, you haven't seen that much um in all of the promotion materials. I think they know they need to do it, but of course, ultimately, it's the ex executive that pay the bill, right? So the promotion have to be very focusing on them. But I feel like then for the analysts, is there anything that you can tell us to make sure that the data that we feed into the system is correct? How to troubleshoot the data sets? If someone proposed two contradicting metrics or formulas, how do we use Tableau to you know harmonize everything? um yeah
00:31:56
Speaker
this is not going to go away. like Big companies are going to have datasets multiple data sources that are not talking to each other. So how do we ah play the analyst role back into the conversation? ah That's something that I am. ah I want to see more. Yeah, yeah that's really that's really interesting. i mean I'll ask you one one one last
AI's Impact on Data Analysis
00:32:15
Speaker
question. So back to your your podcasting project, you kind of ended up building the code to get the data through chat GPT. Do you foresee a world where you would be able to do that task directly through Tableau and have the data populate directly in Tableau? or or and And let me also ask, like that sounds great, but you just mentioned how data's messy and difficult needs to be cleaned. Do you think that even that that should be possible or should be harder for people because then it forces them to be more careful with the data?
00:32:49
Speaker
I think currently everybody ah currently a lot of criticisms received by this AI models are that um it doesn't tell you how confident it is. It always sounded very confident. Always very confident. yeah yeah so when the like I don't want you to always give me a number. I want you to tell me if this data set may be like in ah has some flaws, you need to tell me that today we only receive 50% of the records as usual. But this is the number that we calculated or computed based on the result. But FYI, we're missing 50% of the data. Yeah, yeah, I feel like the FYI part
00:33:30
Speaker
It's never mentioned in any of the results from chat GPT. To be frank, um the Spotify data, I use chat GP, the coding was perfect. But for the other project that I had, I made for iron this. Those are data from IMDB and Wikipedia.
00:33:46
Speaker
chat gbt couldn't solve the coding issue even though he spit up a code that wrote like that was written so organizedly i couldn't run it at all i had to use google ai oh wow you like come on friend let's get one more and then i had to compare the codes one by one so now i am the analyst trying to right harmonize these two systems Right. Because these two systems, even though they will spit out perfect answers, seemingly perfect answers, someone needs to be able to validate and test them. And I think I foresee that to be my role in the future. That's great. Well, I love the future looking piece of this.
Episode Conclusion and Farewell
00:34:23
Speaker
Kevin, thanks so much for for coming on the show. I'll put links to your ah your podcast dashboard and your Tableau public and your site on the on the show notes. So people should ah be sure to check that out. Thanks a lot. I will see you at Tableau conference in a few months and I'll talk to you soon.
00:34:38
Speaker
Yeah, top two soon. It's very nice to be on the podcast. Finally. Thanks for having me. All right. Thanks Thanks to everyone for tuning in to this week's episode. I hope you enjoyed that. I hope you'll take a moment to rate or review this show on your favorite podcast provider. And while you're doing that, you have a moment. If you've checked out any of my books from Better Presentations to Better Data Visualizations to Data Vis in Excel, I hope you'll take a second to rate or review it on your favorite book provider, be it Goodreads, Amazon, or wherever else you get your books.
00:35:12
Speaker
So be sure to stay tuned to this podcast. Hit that subscribe button so you can get more DataViz content. It comes out every other week. And until next time, this has been the PolicyViz podcast. Thanks so much for listening.