Team DataFam Tableau User Group 10 Sep 2026
'Becoming a Tableau Triple Threat' by Nathan Settembrini
1h 10m 10,596 words transcribed
Summary
Nathan Settembrini, co-founder of the data consultancy Allegro Analytics, argues that a world-class Tableau analyst is a "triple threat": strong in business acumen, technical chops and design. Pull any one leg out of that stool and the work fails in a predictable way — a dashboard that is accurate and well-built but ugly gets abandoned; one that is beautiful and exactly what was asked for but wrong destroys trust; and one that is beautiful and accurate but answers the wrong question is shelf ware that gets zero views. He works through each skill in turn: business acumen as partnering rather than order-taking, with a discovery framework built on who, why, what, where, when and how; technical chops covering calculations, interactivity, Tableau's order of operations and a data-modelling toolbox that runs from physical joins to relationships and the date spine; and design, where his single highest-leverage tip is simply more padding. He closes on a worked example — a COVID-era location dashboard taken from sketch to Figma to Tableau — and the argument that these three skills will outlast any BI tool, Tableau included. Hosted by Chris McClellan, with questions from Jen Shepherd, Sarah Burnett, Robert Tullo and Michael McCusker, and an AI-generated quiz to finish.
Listen to this TUG
Prefer to listen? This meeting is also an episode of the Tableau TUG recordings podcast.
Key points
- A triple threat is strong in three things at once: business acumen (what should we build, and why), technical chops (can you actually make it), and a design eye (will anyone want to use it).
- Two out of three is not good enough, and each missing leg fails differently. Ugly but correct gets abandoned. Beautiful but wrong destroys trust. Beautiful and correct but pointed at the wrong question becomes shelf ware — "that gets zero views".
- Learning how your organisation actually makes money is a moat. It is why people in healthcare tend to stay in healthcare: the learning curve is steep, and once you have climbed it, that knowledge keeps others off your territory.
- Old IT habits turn analysts into order-takers pulling tickets — "this guy wants a dashboard, wham, bam" — with no back-and-forth. The fix is curiosity: ask better questions, and keep peeling back the onion.
- The sharpening metaphor for stakeholder pushback: if they set 100% of the requirements you sharpen from one side and get a chisel. Push back and engage and you sharpen from both sides, and get a scalpel — far closer to the optimal solution.
- Challenge without support does not work. Nathan describes being on the receiving end of a wall of nos — "stiff arm, stiff arm, stiff arm" — and concludes you have to earn the right to challenge somebody.
- Discovery runs on six questions in a deliberate order: who is the audience, why are we building it, what metrics matter, where does the data come from and where is it consumed, when is it viewed and refreshed, and how will users reach it. Then wireframe — paper, PowerPoint, Figma or Miro, whichever is fastest for you and your stakeholder.
- Tableau's order of operations is genuinely useful, not just trivia. Filter a Top 10 and it becomes a Top 3 because Top N runs before dimension filters; move those filters to context and the Top N happens after.
- Data modelling is just a toolbox for shaping data so Tableau can do what you want — physical joins and unions on one side, relationships and the noodle on the other. Avoid custom SQL per workbook, or a hundred dashboards later you have a hundred copies of the same query to maintain.
- The date spine (scaffolding, imputation) lets you relate otherwise-unconnected measures on shared dimensions like month and warehouse, and drive all your filters from the spine, instead of flattening everything into one wide table.
- The simplest change that improves every dashboard is padding. Tableau defaults to 8 around the dashboard and 4 between objects; Nathan suggests at least 30 outside, around 20 between boxes and at least 10 interior.
- Design borrowed from the aesthetics of joy: harmony through balance and rhythm, energy through pops of colour, freedom through negative space, and play through circles — which is why bubble charts make people happy and why rounded corners were worth the wait.
- AI mockup tools produce output that looks good and means nothing. It is worth asking what the chart is actually telling you before trusting it.
- These skills outlast the tool. Business, technical and design competence will serve you beyond any BI product, Tableau included.
Jump to a chapter (58)
- 0:00 Welcome to Team DataFam TUG meeting 91
- 0:37 Introducing Nathan Settembrini
- 0:54 Nathan takes over: slides, a Miro board, no Tableau
- 1:56 Who is Nathan: Allegro Analytics and ten years of Tableau
- 3:10 A quick plug for the podcast
- 3:29 What "triple threat" actually means
- 4:41 The three skills: business, technical, design
- 5:36 Rate yourself: where are you strong, where should you grow?
- 7:09 The room answers: loving design, learning the technical
- 8:48 Pull a leg of the stool: two out of three is not enough
- 10:06 Get all three and that is a home run
- 10:57 Skill #1 — Business acumen
- 11:32 Knowing the business is a moat
- 11:57 Partnering with stakeholders
- 12:06 Stop being an order taker
- 13:18 Chisel or scalpel: what pushing back is worth
- 14:31 Making every project a success
- 15:41 The project life cycle, and the two ends we neglect
- 16:06 A simple discovery framework
- 16:44 Who, why, what, where, when, how
- 19:41 Wireframing: paper, PowerPoint, Figma, Miro
- 21:29 Q&A — advice for consultants who cannot push back
- 25:05 Earn the right to challenge somebody
- 25:44 Skill #2 — Technical chops
- 26:50 Calculations: row vs aggregate, LODs, table calcs
- 27:32 Building flexible interactivity
- 28:11 Order of operations: why your Top 10 becomes a Top 3
- 29:16 The data modelling toolbox
- 31:08 Logical and physical layers in the data pane
- 32:24 Why not to write custom SQL per workbook
- 33:25 Joins zipper, unions stack
- 34:48 The stack hack
- 34:59 Relationships and the noodle
- 36:35 The date spine
- 39:41 When separate data sources are still the answer
- 40:51 Skill #3 — Design eye
- 41:20 Visualization best practices and aesthetics
- 42:00 The simplest fix: just add padding
- 42:39 Intentional use of colour, and the screaming cat
- 43:41 The aesthetics of joy
- 44:57 Harmony, energy and freedom
- 46:14 Abundance and play: why we love circles
- 47:17 A real project: the COVID location dashboard
- 48:17 From sketch to medium-fidelity Figma
- 49:43 Adding energy, then freedom
- 51:02 The Tableau build, and 50,000 views
- 52:26 Closing: these skills outlast any BI tool
- 53:25 Where to find Nathan, and thanks to the DataFam
- 54:35 Q&A — low-effort alternatives to PowerPoint wireframes
- 57:15 Careful with AI mockups: good-looking and meaningless
- 58:21 Just draw it on a whiteboard
- 58:37 Build the hex map on the glass wall
- 59:19 Q&A — how far does the final build deviate from the wireframe?
- 1:05:00 Quiz: where everyone is from, and what they do
- 1:05:46 Quiz questions: regex, context filters, parameters, extracts
- 1:07:48 What to present next, and who to invite
- 1:08:18 The leaderboard and the winners
- 1:09:04 Next meeting, the channel, and wrap-up
Transcript
0:00 Chris McClellan: So welcome to Team DataFam Tableau User Group meeting number 91.
0:05 This TUG aims to capture all the in-person and unrecorded presentations and give them a wider audience. So I don't really mind where you've presented before. I actually don't care if you've never presented before. Um, as Anne said, it's a safe space. Everyone's welcome. We've got meetings all the time. We just need a, a time slot that suits you, because you're presenting, and me, and we get that meeting happening.
0:28 So if you want to present or if you've got any questions, please email me at team.DataFam@Gmail.com, and I can walk you through that process as well.
0:37 But today we've got Nathan. The subject is becoming a Tableau triple threat.
0:43 Uh, sounds very interesting, talking about data professionalism, business acumen. I don't know. Nathan's the expert, so I'll hand it over to him, and he can take it from here.
0:54 Nathan Settembrini: Wow.
0:55 Well, thank you very much. I appreciate the, uh, attempt of the last name that you didn't get.
1:05 Uh-
1:05 Chris McClellan: I, I could try yours if you can try mine.
1:08 Nathan Settembrini: Yeah, no, it's okay.
1:09 Uh, hey, guys. I'm Nathan Settembrini. It's gonna be a bunch of slides, and then we'll probably go to a Miro board. We're not opening Tableau at all today. Ooh. Oh, I can zoom. I did not mean to do that. All right. Cool.
1:24 Um, and Chris, feel free to stop me if, uh… 'cause I've only got one screen here.
1:31 Um, you know, if there's questions in the chat. Um, you know, and guys, actually, why don't… if you have a question, um, ask it now.
1:41 I'm just kidding. Uh, but you can, uh, come off of mute and stop me, and I'll try to, I'll try to listen for you.
1:49 Uh, or we can also just do it at the end. I don't think you'll have a lot of questions, but we shall see. Okay.
1:56 So who am I? Um, I'm Nathan Settembrini. I'm the co-founder of Allegro Analytics. We are a data consulting firm. We help companies make better decisions faster by building technology tools that they trust and that they actually want to use.
2:15 And I have been using Tableau, uh, for the last 10-plus years.
2:21 I've been to six Tableau conferences. I'm going to my third Dreamforce next week.
2:28 Uh, if you're gonna be there, would love to meet up and connect.
2:32 Um, you know, I've led Tableau user groups in different places that I've lived, uh, and that's my business partner, Melanie, uh, in the picture over there.
2:42 And been certified a bunch of times. And this presentation is a distillation. It's a combination of a bunch of different presentations I've given over the years to different user groups, uh, internal and external, recorded and not recorded.
2:59 Um, and, uh, I hope it's helpful to folks. It's, uh, it's about being a Tableau triple threat. And actually, hold on.
3:10 Quick plug for the podcast. Uh, I also have a podcast. It's been on ice for the last six months, but we're bringing it back to life.
3:18 Uh, it's a true crime themed data and business podcast.
3:24 Um, there you go. It's on YouTube and all the other things.
3:28 But, um, today we're talking about being a triple threat, and what this means, like, in the dictionary, as you can see, is someone who is highly skilled in three different areas, uh, as it pertains to, like, a sport, um, or in, uh, like performing arts, someone who can act, dance, and sing. In, you know, football, someone who can run, throw, and kick.
3:55 Uh, in Tableau, the way I used to define this actually was, uh, someone who basically, uh, it was more product oriented, and so it was like, oh, you gotta know Tableau Desktop, you gotta know Tableau Prep, and you gotta know, uh, Tableau Server or Tableau Cloud. Um, and I would basically go through each one and kinda… It was more designed for beginners.
4:19 Um, but as I've kind of let it simmer, I've gotten a little less product oriented and a little bit more skill set oriented that's can be genericized outside of purely Tableau.
4:34 But I think, um, these are the things that do make you successful as a Tableau analyst. And so if you wanna be world-class at this thing that we call data analysis, um, these are the three things that I would submit you must be excellent at. Um, business.
4:55 So business acumen, technical chops, and a design eye, let's say.
5:02 Um, and so these are the three legs of the stool.
5:06 Uh, and so business, this is like defining what do we need to build?
5:11 Why should we even be building it?
5:14 Um, technology is, you know, this is the, the ability to drag, drag and drop, the ability to actually make the thing come to life in the tool.
5:24 Um, and design is, you know, make it pretty and easy to use, which translates to delighting our users through the aesthetics of what we've built.
5:36 So let's take a quick pause and reflect on your own skill set. So you could imagine a spider chart or, you know, a little bar chart on your Tableau Public resume, right? That's like business, technical, design. Where would you rank yourself on a continuum in those three areas?
5:58 Um, and then I would actually love for maybe one or two or three… people to share their own, like, self-assessment. Like, what are you best at? Where should you grow?
6:12 And maybe introduce yourself with your name and then say.
6:16 Is anybody brave enough?
6:18 I'll go.
6:19 Chris McClellan: Mm.
6:20 Nathan Settembrini: Uh, my name's Nathan.
6:22 I'm in Jacksonville, Florida, United States.
6:25 Uh, I va- actually value design probably above all else. I'm Italian, you know, I like things to look good. I don't know.
6:35 Um, I'm probably best at the business acumen piece. Uh, started, I'm an engineer, got an MBA, like, kind of leaned hard into the business side. Um, and then sometimes the technical side, you know, I'm, like, pretty technical, but I'm not, like, you know, I'm not a data dev. So, there you go.
6:56 Anyone else brave enough?
6:58 I'll give you five seconds.
7:00 Chris McClellan: I'm, I'm just trying to push people in the chat. I'm sure Sarah can unmute and say something.
7:04 Nathan Settembrini: I think Kelsey Erke should go.
7:08 Or Sarah Burnett, there you go.
7:09 Sarah Burnett: Yeah, I can. So yeah, I really love, uh, the business and design side, so love everything design.
7:17 But on the business side, I like getting, like, the value out of what they're trying to do, which is a bit technical. It's like trying to relate from a business perspective into technical. But the reason I say I'm more, I could do better on technical is 'cause some of it just goes straight over my head, which I'm very happy for all the LLMs that I use these days, particularly when I've gone down a Tableau Next rabbit hole, which I am down at the moment, trying to get it to do things, very technical, which it doesn't want to do.
7:50 Nathan Settembrini: Oh, bless you.
7:51 Uh, Tableau Next is a lot of fun.
7:55 Um, looks like we got some people in the chat. That's cool.
8:01 Uh, Michael McCusker's ability to talk through a very technical solution in a very logical, sequential way with clarity is unmatched.
8:13 Unmatched.
8:14 Um, Kelsey Erke's design-
8:17 Chris McClellan: He's too kind
8:17 Nathan Settembrini: … 10 out of 10.
8:21 Chris McClellan: Sorry, Michael.
8:21 Nathan Settembrini: All right.
8:21 Chris McClellan: You said what? It's a, it's a what?
8:24 Michael McCusker: He's too kind.
8:28 Nathan Settembrini: All right.
8:29 Chris McClellan: Any, any-
8:29 Nathan Settembrini: We'll, we'll forge ahead. Uh, if you-
8:31 Chris McClellan: Oh, sorry, I thought you were trying to get Kate to say something. Maybe she's not in a position to unmute. It's okay.
8:37 Nathan Settembrini: Yeah.
8:38 I think we'll just… Yeah.
8:39 Chris McClellan: That's good.
8:39 Nathan Settembrini: If folks didn't get a chance to share, drop it in the chat. I'd love to see what you have to say once we kind of wrap things up.
8:47 Um, okay.
8:48 So, here's the old pull a leg of the stool act, right? So, um, two out of the three is not good enough. And so if you have a dashboard that is the right thing for the right user, gives them all the information they need, it's technically accurate, but it is ugly, um, and it's hard for users to understand. They show up and they're like, "I don't understand what I should be under- learning from this dashboard.
9:16 How, how do I take action from this?" Then you have failed, right?
9:21 Um, design is an important leg of the stool.
9:23 Um, technical, super important, 'cause if it's beautiful and it's exactly what they wanted, but it is completely wrong, uh, you lose the trust of your users. Um, and if the business leg is gone, uh, you either misunderstood your users, you misunderstood the metrics, misunderstood the data, you built the wrong, useless thing that nobody needs, and you will be relegated to shelf ware.
9:54 Um, so we do not want to build the wrong tool that is technically accurate and beautiful, 'cause that gets zero views, uh, in our analytics. Uh, here we go. So we want all three.
10:10 Um, so we build the right tool. It's built elegantly, it's fast, it's accurate, and the users love what they see, and they actually want to engage with it. That is a home run, baby.
10:26 So, what we're gonna do over the next, I don't know, 30 to 45 minutes, is, um, go through each skill and we're gonna talk a little bit deeper about them, and we're going to kind of go one double click down on a particular area. We cannot give you an MBA in the next 45 minutes. We cannot teach you everything you need to know about Tableau.
10:52 Um, but we can dive deeper in a couple key areas. So, let's start with business acumen.
11:00 So this is all about partnering with our stakeholders so that we get the right tools in their hands.
11:06 Uh, so what are we talking about when we say business acumen? So this is all about learning how your business or organization makes money. What's… How do you measure the impact that is made? What about in your functional area? Like, if you're in HR, you need to lear- learn about people analytics metrics, that type of thing.
11:28 Um, so it's all about diving deep into your c- y- your context. Um, this is why people who, um, are in healthcare will often stay in healthcare, because that learning curve to understand that business is pretty steep.
11:45 And so once you have those skills, that's super valuable, and it actually acts as a moat to keep others from, uh, treading on your territory. So, um,
11:57 the next thing I would say is partnering with stakeholders. So this is not being a- Um, just an order taker, right? So back in the day, we used to, in order to get the data you needed, you had to submit a ticket to IT.
12:15 They would run a query, they would give you a CSV, and, you know, it was very, like, transactional. And so a lot of that kind of old IT mentality sometimes creeps its way into the work that we do as analysts. And so we become just kind of like order takers, pulling tickets. "Oh, this guy wants a dashboard.
12:40 I'm gonna build it," wham, bam.
12:42 And, um, there's no back and forth, right? And so we, we need to develop curiosity for, like, how do we ask really good questions? Why, why, why? Peel back the onion.
12:56 We need to develop our communication skills so that we are consistent and proactive and, like, engaging our stakeholders and letting them know that, "Hey, um, you can expect the next iteration in two days," that type of thing.
13:13 Um, and then this one's my favorite, pushing back, uh, which I kind of already talked about. But the, the metaphor that I like to think about is, um, let's say you have you on this side, your stakeholder on this other side, right? And your goal is to sharpen this object.
13:36 And if the stakeholder is setting 100% of the requirements, and they're just like, you're gonna give them 100% of what they want, then they've sharpened this object down into a blunt instrument, which is a chisel, right?
13:52 If in that process you are pushing back and also engaging with that stakeholder, you now are sharpening this from both sides, and you end up with a scalpel that's a much more precise instrument. It's much closer to what is the optimal solution.
14:13 The same way, like, if you were sharpening from your side without the stakeholder involved, you're gonna create a blunt instrument too, right? It's gonna be just a chisel the other side. And so ideally, you're in partnership with your stakeholder and creating the best possible solution.
14:30 Um, then thinking about, like, how do you make every project a success?
14:36 Well, it starts with that discovery, engaging the stakeholder, whiteboarding, figuring out what should be built, um, working with… Potentially you have a data engineering team that you're partnered up with that you need to deliver requirements to them so that the project's a su- a success, and that's a whole different skill set than working with executives.
14:57 Um, and then the thing that everybody forgets at the end of a project, which is, how do we make sure that this actually gets used?
15:07 Um, and so everything from how you launch the thing and, "Hey, everybody, this is happening," a big splash, and then office hours and, um, sitting with people and showing them how to use it and all that kind of stuff. Um, driving adoption, uh, will also give you ideas to iterate for the next version, uh, and make it even better than it was. So there are millions of books written about business stuff, but those are just kind of a few things that I'd say, um, are really important.
15:42 And so as you think about, like, a project life cycle, um, you know, a lot of us spend a lot of our time in this development area.
15:51 Um, maybe some of the designers here spend a lot of time, like, kind of wireframing and making it beautiful.
15:56 A lot of times we neglect the two ends, the figuring out what should we build, and then the making sure that it gets used and understood. Um, so I wanna take a quick dive into a framework that my business partner came up with that I'm sure lots of people have used in some way, shape, or form. But I really like it because it's so simple.
16:21 Um, and so for discovery, like, this is our goal.
16:27 We wanna clar- clarify the requirements, define our metrics, create some visual representation of what we're going to build, get everybody to say, "Yeah, that's it," and then, you know, take another pass, make it even better, right?
16:41 And so how do you do that?
16:43 Uh, we would recommend you take the basic questions, uh, who, why, what, where, when, how. Uh, and it's in a slightly different order intentionally. So we'll walk through each of them very fast, uh, 'cause I don't want everyone falling asleep, especially the people in America.
17:02 All right, so who? So this is understanding who your audience is. You may have a primary audience, you may have a secondary audience.
17:10 You would design different things for the executive versus the analyst, right? So the who matters.
17:18 Why?
17:19 So this is what is it that we're trying to solve for? Like, what are, what information are we trying to surface? What action are we trying to enable?
17:28 Um, and that's important.
17:32 What? So this is kind of defining the key metrics that lead to that action. Uh, so what, what data, what metrics do we need? Um, and so, you know, key metrics sliced by something.
17:45 Um, how are… What's the timeframe? How are we relating it to, uh, either a goal or some prior period? Stuff like that. Um, and also from a data perspective, what is a row in the data for what we need in order to build what we need to build?
18:06 Um-
18:08 Where.
18:09 Uh, so this is kind of getting to, like, where is the data.
18:13 So on, on the kind of under the hood, where is all this data coming from? How does it come together? And then on the far end, the s- piece of glass that your user is sitting in front of, um, what is that? Right? Is it a mobile device? Is it a big screen? Is it a little screen?
18:31 Uh, where is this being consumed? And then also this is where you would consider accessibility, like color blindness, stuff like that.
18:40 Uh, when.
18:42 So I mentioned time already, but this speaks more to, uh, when will this thing get viewed and how often. So you can think about data refresh frequency, um, how often are we refreshing that extract, stuff like that.
19:00 And then how.
19:02 So how will users access the dashboard?
19:06 Oh, I guess I… Okay.
19:08 Well, anyway, so you can consider the users, uh, what, how they're viewing it, what screen.
19:18 Um, you also wanna think about, uh, security and, you know, where will this dashboard live.
19:26 Um, is it gonna be in some project where the access is locked to the project, or is it something we wanna do row-level security on?
19:36 Um, so you need to think about security as well.
19:41 Um, so those are your questions, and based on the answers to those questions, you can start to draw a wireframe, uh, which could be something as simple as pen and paper, PowerPoint, or an online whiteboard or tool like Figma.
20:00 Um, you get a little bit more pixel perfect.
20:04 Um, anything is better than nothing, um, and sometimes it's just what's gonna be the fastest for you and your stakeholder to, like, bring something together. So usually if I'm with a stakeholder in person, it's a whiteboard or it's a piece of paper in front of us. Um, if it's virtual, it's almost always better to do something like a Miro, um, where you can throw stickies and draw little boxes.
20:33 I don't do anything more complicated than stickies and boxes, so, um, but highly recommend drawing things out, and sometimes it's better for the data to not be right intentionally. You say, "Hey, this is…
20:46 These are, you know, we're putting number signs on there, or we're just making up numbers. This is not intended to be accurate," uh, caveat.
20:56 Um, oh, and the cool thing about doing it digitally is you can whip something up in Superstore and paste actual, like, visualizations that are not the real data, but it shows them what a bar chart with ten members looks like, uh, on the page.
21:18 Okay.
21:19 Skill number two, technical chops.
21:23 Uh, I'll pause here. Were there any questions or anybody wanna talk about anything?
21:29 Chris McClellan: There was. Uh, Jen Shepherd has a question if she wants to-
21:32 Nathan Settembrini: Oh, I love Jen Shepherd
21:33 Chris McClellan: …
21:34 unmute.
21:34 Jen Shepherd: Hey, Nathan. How's it going?
21:35 Nathan Settembrini: Hey, Jen. How are you?
21:37 Jen Shepherd: Good. So yes, my question is wondering if you have advice specifically for consultants, although not exclusively. If you maybe just have a short amount of time to work with them and they're asking for things that aren't best practice or they're inefficient or you know it's not a good idea, but you're their consultant, so you need that business, or maybe they just aren't used to being questioned.
21:58 It just seems a lot harder to have that kind of back and forth in those settings. Do you have any advice for that situation?
22:04 Nathan Settembrini: Hmm. Yeah, I mean, there's definitely a kind of an emotional intelligence component to knowing when… like, what type of personality you're dealing with and do they… Like, some people love to get pushed back against, and you would, like, you would think, uh, that, you know, hey, I'm challenging them. Are they gonna, are they gonna be mad at me?
22:30 Like, I don't wanna hurt their feelings, but, like, pie chart's really not the answer right here. Like, um, I think some of it is confidence in your own skills and your own expertise, and, and then sometimes it's the communication of, uh, how you say it. And, you know, you can reference visual best practices, right? So you can say, um, "Oh, that's, that's a really good idea," you know, so kind of a crap sandwich, but, "That's a really good idea to build, uh, to have that, those numbers there.
23:07 Um, you know, I've been studying visual best practices for a long time, and, and one other idea could be this," or, you know, like, um, that kind of thing. Or you kind of frame it in either from their perspective or from their user's perspective to let them know the downside.
23:26 So if you say, like, "Hey, if we add ten more filters to this dashboard, um, we can do that. There may be a performance hit because of that. Um, you know, it could, instead of loading in three seconds, it could be ten seconds. Is that, is that okay?" Like, and they're like, "Oh, no." You know.
23:49 Um, and then, you know, you can push back, and then if you get to the point where they just, they're like, "No, it's, it's gotta be this," then kind of knowing when to relent and, um, just say, "Okay. Well, that's fine. We'll, we'll do that." And, uh, yeah.
24:08 I don't know. Is that kind of the direction of your question, Jen?
24:13 Jen Shepherd: Yes. Thank you.
24:14 Nathan Settembrini: Yeah.
24:17 Um, yeah, I mean,
24:20 that's the, uh, kind of the art and the science of, um, partnering with stakeholders. And, you know, I, I'd recently dealt with a, um,
24:34 let's just… It, it's not a contractor, uh, it's like somebody who's working with us on not Tableau stuff, uh, so it's no one that you guys know. But, uh, we were designing a little PDF thing, and I, I was like: "Hey, I'd like this changed, this changed, this changed." And she came back with a big old brick wall of nos to the point that actually sh- that shut me down super hard.
25:03 It was not… There was no, um, like, feeling like I'm heard. It was all just, like, stiff arm, stiff arm, stiff arm.
25:12 Um, where maybe I kind of just wanted her to just say, like: "Okay, I'll do that, I'll do that, do that." Um, and so that's kind of ex- the counter example of, like, where if you just come back with all challenge and no support, you know, there's kind of this, like, support/challenge matrix of, like, you have to earn the right to challenge somebody.
25:34 Um, so yeah.
25:36 There you go.
25:38 All right.
25:39 Uh, thank you for the question, I love that. Um,
25:44 okay.
25:44 Skill number two, technical chops. Probably something that most of you are stronger than me at, but, um, you know, I passed the Tableau Consultant Certification. That was not easy.
26:00 Uh, so I know my stuff. But, uh, a lot of times in the real world, in big corporate building dashboards for clients, uh, you don't get to do a lot of the cool, super technical hacky stuff.
26:15 Um, you end up kind of doing a lot of bar charts and, you know, more functional things.
26:22 But, uh, I wanted to walk through, you know, just if somebody is watching this that doesn't know anything about Tableau or they're just getting into the game, I want to kind of bring them along. So experts stick with me. Um, but our goal here is to build elegant solutions that are accurate, accurate, timely, and performant, which means that they don't take 40 seconds to load.
26:48 Um, okay. So in order to do that, we need to learn about all the different types of calculations that we'd need to run, uh, or build, right?
26:58 Um, and, you know, that's everything from is this calc, does it need to happen at the row level or in aggregate?
27:07 Um, kind of on the fly. If it's a ratio, it's probably sum of this over sum of that, right?
27:13 Um, level of detail calcs, table calcs, uh, these are all things that as a Tableau triple threat you need to have in your toolkit.
27:24 Uh, we're not gonna… I'm not gonna teach you any of those things. I'm more just, like, making more the beginners aware of the things that I think you should learn.
27:32 Um, there's a whole category of, like, how do we build flexible interactivity into our dashboards so that the users can explore and that you can get to more than what's shown in the splash experience when you land on the dashboard.
27:48 Um, and so we do that through parameters, different types of actions, dynamic zone visibility, which is probably my favorite feature from the last five years, uh, map layers, which, uh, I'm not smart enough to use in real life, uh, even though, uh, Michael showed me how to do it.
28:09 Um, and then I will say that Tableau's order of operations, understanding when things happen in sequence, is not just, like, good bedtime reading. Like, this is actually useful as you're building certain visualizations to understand why if you're doing a top 10 and then you filter it and it goes from 10 to three, what… It's like, why, why do n- why do I not still have 10 of the remaining resulting?
28:42 Well, it's because top N happens before those, um, dimensional filters, right? And so if you instead move those dimensional filters to context, then your top N will happen after the filter. Uh, and so yeah.
29:03 Let's see.
29:04 Um, so yeah. I mean, there's all kinds of great examples of why the order of operations can be used or how it can be used to your advantage, but highly recommend digging into those. Um, and data modeling. So this is, uh, what we're gonna dig into.
29:22 Um, I developed this super data modeling toolbox, uh, which is basically just categorizing all the different ways that you can shape your data for analysis.
29:38 Um, that's pretty much all data modeling is. It's like crafting the data to enable Tableau to do the thing that you want it to do. Um, and not…
29:50 It's, it's kind of like a toolbox because you wouldn't use a hammer to screw in a screw, and you wouldn't use a tape measure to…
29:59 I don't know.
30:01 My analogy's falling apart. But, um,
30:06 I'm gonna, I'm gonna fly through some of this because some of this is very, uh, obvious to a lot of you. But, um,
30:16 let's see. So intro to data modeling.
30:20 We're, we're actually not gonna get into, like, all of the nitty-gritty. What I do wanna show you is I wanna highlight a couple important things. Um, and so we're gonna talk about the, uh, the different types of data connections that you could use, uh, when you're building a data source in Tableau, um, or upstream from Tableau.
30:44 Um, and so, you know, joins, unions, relationships, um, you know, separate data sources, which means maybe you're gonna blend and someone's gonna smack you on the hand.
30:56 Um, and then some, like, down here at the bottom we have some kind of alternate, uh, versions of, um, using the union or relationship. And we'll t- that, those are the ones that I wanna highlight. But, um, I think understanding the logical layer in Tableau, so, like, when you go to that data pane, you connect to a data source, and you drag something onto the canvas, what you see first at the top layer is this logical layer.
31:26 If you drag another table in, then you'll get the noodle connection, and that's a relationship. If you double-click down into it, then you can go down to the physical layer where tables can actually be physically joined, merged together, um, and you have, you know, all kinds of risks of Cartesian joins and, you know, stuff getting dropped that didn't match the, the join criteria, stuff like that.
31:51 But, um, I think it's helpful to at least understand that these two layers exist and that you would use them in different ways, uh, to accomplish different things.
32:02 And so back to our toolbox, um, all these pink ones on the left-hand side are the physical joins and unions.
32:11 Um, those can actually happen in your data warehouse if you want.
32:15 They don't have to happen in the Tableau data pane, and actually it's probably best practice for them to not happen there unless there's a reason to do so.
32:25 Um, and ideally, you're not writing custom SQL for every workbook, 'cause then now, you know, a hundred dashboards later, you've got a hundred versions of the SQL code. Need to change something, that's a problem.
32:41 Uh, so highly recommend against doing that.
32:44 Um, but the things on the right in, in yellow are things that you can really only do in Tableau, um, using their relationships, the noodle.
32:57 Um, super powerful, but also has some gotchas.
33:02 Um, I'm not gonna walk through each one of these, but, uh, physical join, just bringing two tables together.
33:12 You know, there's different types of joins, and it has to do with what you're keeping. Uh, this is kind of like doing a VLOOKUP almost, um, but super powerful. Uh, union, this is, if joins is like zippering the data up this way, unions is stacking the data like this.
33:33 Uh, and so it finds the common column headers and combines the data that way.
33:40 Um, and this is the one that we can use to our advantage. If we have a situation where, let's say we're trying to build a chart where, um, we wanna show two different things on the same chart. They have some attributes that are common and some that are not.
34:02 Um, and maybe one of our common fields is a date.
34:06 And, uh, so this example here, headcount in seats. This is from my corporate real estate days. Um, and so this is, we wanna track how many people are assigned to seats over time. And, um, so we basically have a lot of common dimensions and, um, you know, heads and seats are kind of two different measures.
34:31 And, but we wanna show them all on the same chart. This is how we used to, like, bring those things together back in the day before we had the noodle.
34:40 Um, but it's still useful sometimes, um, to union these type of things together. Um, and so when you have a lot of important stuff in common, but then just some things that aren't, uh, this is the stack hack, as I'm dubbing it.
35:00 All right, so moving into the logical world. So relationships, this is our noodle.
35:05 Things brought together through a, a relationship clause.
35:10 This region in this table is the same thing as region in this other table.
35:15 Uh, the difference here is that in Tableau is not, um, physically bringing these things together in the data source earlier in the order of operations. It actually brings the things together later.
35:29 Um, as it's computing the viz live, it's running SQL that's based on what's in the viz, in the viz and how things are related.
35:39 So you can end up with some weird stuff happening if you try to, um…
35:46 You have to think about, like, what does Tableau know, uh, from this table that's related to that table?
35:53 Um, and if, if it has no way to relate people to returns, you know, like for the data to get all the way around from orders to returns, um, or if the, you know, in that path you end up with a bunch of things that, um, are, like, at a higher level. You could end up with, like, a big, really big number getting pulled through because it just doesn't have a way to make that connection.
36:20 Or it's, in this case, it's doing it at the region level, and so every person would get the sum of the returns for their region, uh, which would be a really big number, not the sum of the returns for that person. So, um, really important.
36:36 One of my favorite hacks, uh, that can also be a booger if used incorrectly or, uh, in a way that's too complicated, um, is borrowed from data engineering. There's this concept of a date spine.
36:51 Um, so date spine is basically just, uh, you know, a lot of times in the Tableau world they'll talk about, uh, scaffolding or imputation. It's basically the idea of, like, forcing something into the data that doesn't exist. Um, so if we have a date spine, let's say it, it's basically just at whatever grain we want, so let's say it's daily. Um, for the last, say our company's been in, in business for 10 years, we would go back 10 years, have every day, um, for the last 10 years and maybe the next 10 years or something like that.
37:26 And, um, then let's say we have a bunch of different effectively data sources that are all related only by date and maybe one other thing like region.
37:40 Um, and so we wanna know… Or let's say in this ca- this is, like, a, um, supply chain example.
37:47 So let's say we have warehouse and date.
37:53 And actually for simplicity let's make it a month, so the beginning of the month, right? So we wanna know by month for each warehouse how, you know, how many receipts did we get, how many returns, how many sales, right? We wanna be able to show all these kind of unrelated things on the same dashboard, same chart, same set of filters. Uh, we can use this d- dimensional spine.
38:18 And so basically for every month, so January, we would have all, let's say there's 10 warehouses, we'd have, uh, 10, 10 Januaries and o- one row for each warehouse. And then February, one row for each warehouse. And so the date spine is basically a Cartesian join between the different dimensions that you wanna use that a- that are logical.
38:40 Um, so, uh, super interesting way to, um, to use the relationship model because you could basically pull all of your filters from the date spine.
38:57 Um, and then when you're charting the, you know, you'll throw a month from the date spine onto, um, columns for example.
39:08 And
39:10 then our measures that we wanna show are all gonna exist down here.
39:15 Um, and so you basically can pick and choose from whatever measures. And so if you wanna know what percent of, uh, total returns over total sales for this month, you have returns and you have sales. And they're, they came from two different places, and rather than building some kind of like huge flat table, we can keep them nice and tidy and just relate them on the things that we care about.
39:40 Super helpful.
39:42 And then sometimes you just gotta do the old-fashioned separate data sources. And you can use parameters instead for date filters, um, 'cause the parameter exists outside of the data source and can filter things across many data sources.
39:59 Um, you can do filters based on related things, but, you know, sometimes that's wonky. It kind of acts as like an inner join.
40:07 But,
40:09 um, so I'll just leave this there for a second. But this is basically, like, when would you use one versus the other?
40:17 When to use the hammer versus the screwdriver?
40:20 Uh, it kind of depends. And so I think just being aware of what all of those options are makes you a better an- analyst, analyst. Um, so there we go.
40:33 All right, I'm gonna pause there.
40:36 Any questions?
40:38 Custom SQL should be banned.
40:42 Chris McClellan: I was gonna read out that, but, um, but yeah, there's a lot of chat. It's good.
40:46 No real questions I can see, but there's a lot of chat going on.
40:50 Nathan Settembrini: Cool.
40:52 All right, so this one, I think I'd like to make this into a whole presentation, um, about design, 'cause I think it's still, uh, kind of underappreciated.
41:05 But, um, this is how do we leverage both, um, like graphic design and data visualization design best practices to make beautiful, intuitive tools that people actually want to use.
41:20 So there are gobs and gobs of great books about data visualization best practices, when to use what kind of charts, how to lay out your dashboard, um, how to think about accessibility.
41:35 Um, and then, you know, from the graphic design and UX/UI world, we can get a lot of stuff about aesthetics, you know, having a hierarchy of different typography. You know, like even just having th- like picking three sizes that you're gonna use on a dashboard, uh, and sticking with that is super, can like really help it feel less like, ah, all over the place.
42:01 Um, I think the m- probably the most impactful, simplest thing that everyone could do tomorrow to make all their dashboards better is padding.
42:11 Um, adding more padding. W- way more than you think. You know, Tableau forces you to do, or starts with eight by default, right, around the dashboard.
42:20 And you drag something else in, it puts four, right?
42:23 Uh, I would submit at least 30 on the outside, maybe 20 between boxes, and, you know, at least 10 interior. Um, that'll just give you so much more breathing room.
42:39 Um, use of color, uh- Uh, there's, you know, lots of people's perspectives on, on this. But, um, you know the old cat emoji. What, what is that one? Um…
42:56 Chris McClellan: Yeah. The screaming cat.
42:57 Nathan Settembrini: The screaming cat. Yes, from… What's his name? Steve?
43:01 Chris McClellan: Steve Wexler.
43:03 Nathan Settembrini: Sorry, Wexler. Yeah.
43:04 Chris McClellan: The, the book of dashboard.
43:05 Nathan Settembrini: He's the best.
43:06 Sorry, Steve, if you ever see this. You are the best. Okay.
43:11 Um, but yeah. Intentional use of color, really important. Um, and, you know, it's basically like how can our design choices, both in color, charts we use, everything is intentional, um, and it has some consistency to it, and everything has some meaning to it. So that sounds a little maybe obvious, but people don't use it in practice.
43:41 Um, this book is awesome.
43:45 Uh, it's not a dashboard book, but, um, it's actually all of her examples are more like design, like physical spaces and things like that. But, um, it's all about the aesthetics of joy and how, um…
44:03 She's got like 10 different ways that you can infuse joy into your designs. Uh, and so Robert Crocker and I, if anyone remembers him, uh, he's kind of moved on to software development, but, um, he was a big design guy in the Tableau space maybe five, six years ago.
44:25 Um, he and I did a presentation together about infusing joyful design into your Tableau dashboards, and so that's where some of this comes from.
44:34 Um, I'm not gonna go through all 10. The book is really incredible, and you could, you know, pick it up on Amazon. But, um, so here's just a few examples, and then if we have time, which I think we do, um, I can go through how this comes to life in a, in a dashboard. But let's just talk, uh, about these in abstract first. So harmony.
45:01 So this, I mean, as you can see from the photos, this is all about balance.
45:06 So visual balance, um, using rhythm, symmetry, geometric patterns.
45:14 Um, you know, there's, there's something like oddly satisfying about, you know, these, these, uh, geometric patterns with some sort of rhythm to them.
45:25 Um, energy. So this is all about using bright pops of color, um, to infuse joy into your dashboards.
45:34 Freedom. You know, we, I already beat this horse, but, um, the padding. So this is all about open spaces, um, using negative space and natural textures.
45:50 So like, um, it's the reason why in a lot of design, like physical spaces, you'll see, um, the use of natural woods and plants and stuff like that. It, it gives you this sense of, um, kind of wild nature.
46:06 And then like big tall ceilings and, um, all that is like kind of speaking to this freedom side.
46:15 Abundance. So this is why I think why bubble charts make us so happy. Uh, this one combined with the next one, um, which is, you know, just all about quantity and variety and like lots of interesting, uh, textures and patterns.
46:32 Um, and then play. So she goes on and on about the circle, um, and just why as humans we love the circle. Um, and I, I can't quite remember, uh, why that is, but, you know, just remember people love circles. So there you go.
46:52 Um, and so this is like, you know, adding curves and, and things. You know, we finally got rounded edges, like, you know, that's… There's a reason for all this. So, um, okay. So I'm gonna try to flip over to share a different screen.
47:12 I'm gonna change the aspect ratio for y'all. There we go.
47:17 All right. So this is a dashboard. This kind of dates this presentation. But, um, during COVID, this big tech company was trying to get information about all the different places in the world where they have employees, get that information about vaccines and cases and etc., etc., like how safe is it, it is. Um, and this dashboard was designed to be embedded in their intranet site.
47:46 Um, and actually, this dashboard is on my Tableau Public if you wanna go play with it at the end. Don't look now.
47:53 That'll ruin it.
47:55 Um, but worked with my client to come up with this sketch. Uh, she actually did the first version of this. Um, and then we kind of walked through and like tried to understand, like what are, what are all these things, and how do we calculate this metric? Where is it coming from? You know, all of our like who, what, where, when, why kind of questions.
48:17 And then from this, we went from a low fidelity dashboard into a medium fidelity dashboard.
48:24 Um, so this I drew in Figma.
48:27 Um, you can see
48:31 there's, you know, it's all kind of black and white right now except for this little line.
48:36 Um, and there's kind of some balance to it, but also, you know, this feels a little off. It's like not quite 50/50, and it's not quite, uh, two thirds, one third. So, you know, it's feeling a little funky.
48:53 Um- But, you know, we're at least starting to try to infuse some harmony into it.
49:00 So, um, you kind of lay things out a little more symmetrically, and there's, like, visual… There's this concept of, like, visual weight, right? Where if you have, like, some really big, like, dark gray thing over here, um, you kind of need something over here to, like, bring it back to center. Um, I'm sure we've all experienced this, where you think you'll be able to lay out a chart or a dashboard a certain way, and then once you actually have the data, you see that, oh, wow, there's, um, this looks different than I imagined, therefore, like, the visual weighting is off and I need to actually change my design.
49:40 Um, but okay, so that's harmony. Then we're gonna take energy. We're gonna infuse some color into it.
49:48 So we did, and it's a little bit of a screaming cat at the moment.
49:53 Um, we want to kind of do good or bad, uh, with these icons. Um, we kind of, kind of… We've got color, but it's not really meaning anything at the moment.
50:06 Um, and so we need to infuse some meaning into that.
50:12 Um, and also it's feeling, it's feeling a little tight.
50:16 Feeling a little… Like, these icons are too big.
50:19 Um, and you might say, "Oh, well, icons can't be small. Like, you won't be able to tell what they are." But, like,
50:27 give yourself a little bit more padding and you will, you will thank me later. Um, so we're gonna add some freedom into this thing, and now it's like, ooh, it's starting to look good. Mm.
50:39 You know, I kind of spread this guy over a little bit so that it's not quite… Like, it was a little too close to 50/50. Uh, felt in- unstable. Um, now it feels a little bit better.
50:50 I could probably get away with… But, you know, we're also trying to line these babies up, right?
50:55 So there's gives and there's takes.
50:59 But, um, and so this was, like, the last Figma dashboard.
51:03 Uh, the beautiful thing about Figma, pixel perfect, right? So we can put things exactly where they were in Figma in the Tableau dashboard. So this is the Tableau version.
51:14 You can see we had to make some, uh, font changes, you know, going with Arial just so that it renders right. Um, and, you know, we had a couple other little things that we had to do. But really came together. And so this dashboard, you basically start here, and you click into your location, and it takes you to the location, and you can zoom out and zoom in, like, to city or county or state.
51:42 Um, yeah, it got, like, I don't know, 50,000 views during COVID, uh, internally. We actually hosted it on Tableau Public as a hidden, uh, thing. That was the way that we got around licensing. Um, but yeah, there you go.
52:02 So that's how we use some of these design principles in a real project.
52:08 Any thoughts, questions, anything so far or we're gonna land the plane here?
52:15 Chris McClellan: Everyone's talking about, uh, rounded corners, visual weight, et cetera. Just reinforcing what you were talking about.
52:22 Nathan Settembrini: Nice.
52:25 Um, yeah. So all that to say, I think, you know, becoming a world-class analyst is a continuous journey. And, um, I think these skills will outlast any BI tool, Tableau included, um, and will serve us, uh, for the long term to be able to s- to serve our stakeholders, our clients, our peers, uh, and the next generation behind us, um, to pass these tools on to them.
53:04 So I would just encourage you, like, you can't go wrong investing in yourself in these different areas if you want to stay in the data technology world.
53:13 Um, so I would highly recommend you kind of taking what we've kind of briefly talked about today and, um,
53:23 continuing to invest in yourself.
53:25 So, uh, if we're not connected on LinkedIn already, I would love to connect with you.
53:31 Um, check out the Data Secrets Podcast. Feel free to shoot me a note.
53:36 Um, you know, we're always looking for amazing data people at Allegro. Um, we're growing really fast and have some really exciting stuff happening in the future, um, in Tableau land and outside. But, um, you know, all this, we're really just here to make an impact. Um, and so, you know, I love the DataFam. I, it…
54:07 You know, everybody who has been a part of it in the past and, like, written blogs and given presentations and, um, you know, just thank you so much to each of you who invest in this community. I think this community is one of the greatest things, um, that I've come across in my career. So, uh, thanks to all of you. And, uh, Chris, thanks for having me.
54:32 And I am done.
54:36 Chris McClellan: Not a problem at all. While you're there, Robert's question was, "Going beyond pencil and paper, my preferred wireframe method, because I'm just too slow at everything else, is PowerPoint sometimes works."
54:47 Nathan Settembrini: Mm-hmm.
54:47 Chris McClellan: "Thoughts on low, low learning effort alternatives."
54:53 Nathan Settembrini: So lower effort than PowerPoint?
54:56 Robert Tullo: Well, um, some, well, PowerPoint-
54:59 Nathan Settembrini: Go ahead
55:00 Robert Tullo: Yeah, PowerPoint, like you said, is kind of like, oh, manipulating. Obviously, it's not always pixel perfect. You, you do gotta strike a- the balance between how much energy do I wanna put in this wireframe and all the iterations than actually just trying to build it. You know, sometimes it's like, I'm gonna get it kind of close, I'm gonna build it in Tableau, see kind of how it goes, and then get the iterative feedback from-
55:24 Nathan Settembrini: Mm-hmm
55:25 Robert Tullo: … you know, r- rather than getting it all done up front and then, you know, implementing. Kind of like, I guess, the difference between a waterfall, uh, software development design and agile. You kind of get it kind of close with the agile and then you iterate and you're, you're kind of happy.
55:42 Nathan Settembrini: Oh, yeah.
55:43 Robert Tullo: Um-
55:44 Nathan Settembrini: Yeah, no, I, I think, um, so there's Miro, Mural, FigJam, any of these kind of tools where you can just throw sticky notes and boxes onto the screen-
55:59 Robert Tullo: Mm-hmm
55:59 Nathan Settembrini: … um, live with your stakeholder.
56:04 Um, you know, so sometimes what I'll start with is I'll take… Okay, if we know we're gonna build a kind of 16 by nine box of content, um, we'll figure out like, okay, well, what are we talking about? Well, okay, we really care about this metric. Call it, you know, sales, right? Sales, and then we also care about, um, I don't know, customer satisfaction, right?
56:35 It's like, well, how much more do you care about sales than customer satisfaction? Well, you know, it's like, you know… So you can like start to take these panes and like figure out how much space on the board do you wanna take up.
56:49 Um, and then you can say, okay, so now within sales, what do we care about? And you can like drag sticky notes. You know, you're not trying to like draw bar charts live with your client there, but, um, yeah, I think any tool w- I think probably the answer is like, whatever tool you're most comfortable with and can move the fastest in, that's probably the right tool.
57:14 Um, there are also all these like, um, like, what's it, mockup.AI, M-O-K-K-U-P.AI, um, is more of a, um, AI-enabled thing. I've, I've seen some outputs from it that are good, some that are kind of like, they look good, but they're not meaningful. You know? It's like, oh, it looks cool, but you actually like figure out, what, well, what is this chart telling me?
57:43 It's like, I don't know, you know?
57:46 Um, so you gotta be careful with the AI stuff 'cause it will, it'll do a lot of stuff that looks good that's not necessarily useful.
57:54 Um, and then like I've, I've coded, like vibed up some, um, some mock-ups in Claude before that actually looked pretty good. And, um, Claude's pretty good at, if you tell it, "Hey, keep everything the same, but like let's, let's shift all these things over a little bit. Change this chart to something else," like, it's better at iterating than some of the other tools, but…
58:22 Chris McClellan: And I've just had it with clients before where you just say, "Just draw it on a whiteboard."
58:27 Nathan Settembrini: Yeah.
58:27 Chris McClellan: You know? Draw, draw it on a whiteboard, send me a photo. I, I need to know what's in your brain.
58:32 Um-
58:32 Nathan Settembrini: Mm-hmm
58:33 Chris McClellan: …
58:34 and, and also probably, you know, a challenge for some of the Aussies as well is in the US when you do a hex map of the states, it's really easy. And in Australia, when you do a hex map of the states, it doesn't quite work properly 'cause we've got one state that's inside another one.
58:49 Um, and I had a client that was saying, "Oh, we, we need a hex map of Australian states." And okay, and you, and you say, "Here's some sticky notes, and there's a, a, a glass wall in one of the offices. Build the hex map on the glass wall.
59:03 Send me a photo when you've got it the way you want it to work." And, um, and it took them about a week to realize what I was trying to tell them was it's pretty hard.
59:11 Um, but they got there. They got the hex map exactly how they wanted it because they'd built it out of sticky notes before we even did anything in Tableau.
59:20 Um, Jacob, did you wanna ask your question?
59:23 Jacob Rothemund: Oh.
59:25 Um, my question.
59:28 So how much do you think that you deviate in your final build, um, from your wireframes? Because, like, yeah, I find with me, I use the wireframe as a guide, and it's definitely not the rule of how my end build will come.
59:43 Nathan Settembrini: Mm-hmm.
59:43 Jacob Rothemund: But I wonder sometimes, you know, may- it, it probably depends on your stakeholder and your client-
59:49 Nathan Settembrini: Yeah
59:49 Jacob Rothemund: … because if they've got a lot of trust in you, you've got a lot more flexibility. You can kind of say, "This is what I'm thinking, but just trust that during the build it might not end up looking like this, but you'll definitely get something better." And yeah, I just, you know…
1:00:07 Wireframes are good to get ideas on paper, um, but I definitely don't use it as my rule and make it match.
1:00:16 Nathan Settembrini: Yeah.
1:00:17 I think it, it depends on how clear you are at the beginning what, what you wanna build. And
1:00:29 yeah. But even then, sometimes you see the data and you realize, um, that something you thought would be like five members on a bar chart is actually like 150. Uh, and you're like, "Okay, we might need a different visual for this," or yeah-
1:00:48 Jacob Rothemund: Is that-
1:00:49 Nathan Settembrini: … something more
1:00:50 Jacob Rothemund: … is that Andy Cotgreave who has that story?
1:00:54 Chris McClellan: Andy Cotgreave's story is exactly that, yes.
1:00:57 Building, building the wireframe and putting the real data in it, and the whole thing just doesn't make sense anymore.
1:01:02 Jacob Rothemund: Yeah.
1:01:02 Nathan Settembrini: Yeah.
1:01:04 Yeah, I've not heard that, but I've experienced that, uh, many times.
1:01:08 Yeah. I– The most important thing of the wireframe is to kind of force you to just be thoughtful about what it is that you're gonna put on the paper. And, um, and also, you know, I've had clients where they came to us with, you know, on, on our operations dashboard, here's all the metrics we want, and it was like two pages of bullet points of like all this random stuff.
1:01:36 And so we went through and like, we had to basically take it to the chopping block and like cluster things together, pick the most important ones, and you kinda like create almost a hierarchy.
1:01:51 It's like, yes, you gave me the entire ocean, but we can't… You can't look at all the whole ocean at once, right? You have to kinda focus on the Baltic Sea first, uh, or something, I don't know. Um, but so we ended up with this kind of like cockpit level view, which was cut into four sections.
1:02:13 Um, and then the idea would be if you see something in the cockpit that you need to pay attention to, then you would go down a level into, let's say, inventory, and go to the inventory dashboard. In the inventory dashboard, you can go a level deeper from there, so.
1:02:31 Chris McClellan: It also depends on the client a lot, right? I, I know, yeah, and, and yeah, Jacob brings up Andy's story. I, I always laugh when Andy tells that because it's, it's the shining example of, of why you never use wireframes because the real data is not the same as the mocked-up data.
1:02:48 Um, and I can remember projects where I've deliberately talked clients out of wireframes, like, "Don't build a wireframe. Let's just do it live in Tableau right now," because they were wanting to spend too long on wireframes.
1:03:02 And obviously, other clients where you say, you know, you need to build a wireframe and, and the same thing. Pick up a piece of paper and some pens, go to a whiteboard and draw it. Don't try to be techy, just try and draw out what you want so that you can see.
1:03:18 But I always think a wireframe should take a little bit of time, not the, you know, the next month to actually design the wireframes perfectly because it's, it's just gonna change, right?
1:03:30 It's always gonna change.
1:03:33 Nathan Settembrini: Yeah.
1:03:35 The other way that we've used wireframes in the past is to like, once you have that, now it becomes kind of your visual checkbox for going, getting that data, right?
1:03:47 Chris McClellan: Yeah.
1:03:47 Nathan Settembrini: Where it's like, okay, well, we have, we have the data for the upper left quadrant and the right quadrant. You know, the bottom left is yellow, and the bottom right is red. You know, so it's like you can start to kind of build it out where it's like under construction. You have your blueprint.
1:04:03 Um, but yeah, I mean, you're probably resizing things depending on what the real data looks like.
1:04:10 Chris McClellan: Yeah, for sure.
1:04:11 Nathan Settembrini: The, the goal should not be to re-reproduce your wireframe perfectly. Your goal should be to let the wireframe guide your de– your development and then let your stakeholder kinda like work with them to iterate to create the best product at the end of the day.
1:04:30 Chris McClellan: Okay. No other comments or questions. Looks like we'll move on.
1:04:34 So again, Nathan's contact details, LinkedIn, Tableau Public, or the company website as well.
1:04:42 It's all in the recording. It's easy to see later on.
1:04:46 Um, and we'll move on to the survey. Has anyone or does anyone need more time to do the survey?
1:04:52 Let's go and check the results.
1:04:56 Uh, quite a few people in there. Um, so where, where do you live?
1:05:00 State and country, just seeing where people are from. So we've got, ooh, three different countries today. Australia, all the way through the US, and New Zealand.
1:05:11 Uh, what industry do you work in? So a lot of different industries there. Television audience measurement, that would be an interesting one. Um, so all, all kinds of data and for all different purposes.
1:05:24 And how many years have you been using Tableau? So a nice spread, but a lot of people have been using it for nine-plus years today, which makes the quiz questions more interesting, right? So all the quiz questions generated by AI, but I've sort of checked the answers, and it was just a, um, a, a nice mix of Tableau knowledge today.
1:05:46 Um, first question is really long and really hard. A view with 5 million rows uses a calculated field containing a regex match function.
1:05:55 Which statement about its execution is correct when the workbook is connected live to a database that does not support regex functions?
1:06:04 You've gotta understand the question first, but the correct answer is that Tableau uses that regex locally after it's retrieved the data from the data source.
1:06:14 So that's Tableau just getting around that sort of, um, functionality where the database doesn't support it, but Tableau can do that.
1:06:23 And the next question is: You have a view with a context filter on region and a regular filter on category. In what order does Tableau apply these filters when rendering the view?
1:06:33 There, there was a lot of conversation in the chat about the order of operations because people had answered this question before Nathan was talking about the order of operations. But it looks like everyone got it right, or everyone that attempted it got it right. And the next question is: Which statement accurately describes the capability of the parameter action in Tableau?
1:06:55 It can set a single scalar value for the target parameter.
1:07:00 Parameters are always a scalar value, right?
1:07:04 And you need to… Or this next question: You need to reduce the size of a Tableau extract file. Which type of filter should you apply?
1:07:14 A data source filter's probably f- right as well, but an extract filter would be the right one.
1:07:21 What does the expression IF NULL SALES 0 return when SALES contains an empty string for a row?
1:07:30 So I put this in there just to maybe sort a few people out. Um, I would say that SALES is always a numeric value, but if it contains an empty string, then it's not a num-num- not a number.
1:07:42 So it would just return an empty string 'cause that's not a null.
1:07:47 Um, that's the end of the quiz. What other questions are… What topic should be presented at a future TUG meeting? So we've got some ideas there of different things that people want to see, and if people wanna present on any of those subjects, let me know. And who should, who would you like to see present at a future TUG meeting? Andy Kriebel is the only response.
1:08:08 Two people said that.
1:08:10 Okay, cool. I need to contact Andy.
1:08:13 He's been on a few other TUGs recently, but, um, but yeah, we'll, we'll get in touch.
1:08:19 And the leaderboard is now based on how quickly you answered the questions and how many you got right.
1:08:26 So Jen Shepherd in a minute and 17 seconds got all of them right, closely followed by Anne and Jacob. And the rest, I don't care if you got them right or wrong because you've learnt something during the day and during the presentation, right?
1:08:40 So that's good.
1:08:41 Maybe not everyone got to the end as well.
1:08:46 Um, but let's go back to here.
1:08:50 So as I said, all those questions are generated by AI, but you can submit your own questions. And if you've got other TUGs, they can use that quiz system as well.
1:09:00 Uh, it's just for everyone to use if they need to.
1:09:04 But our next TUG meeting is next week. Danny's going to present why some visualizations are instantly understood.
1:09:12 Uh, it's at that time, so it's one meeting all those different time zones in the world. But again, would be another great presentation, um, or you can watch the recording on YouTube later.
1:09:24 The same as this meeting, everything we do, everything a lot of other online TUGs do is on the YouTube channel @TeamDataFam.
1:09:32 There's a podcast at Spotify and Apple. I'll have to catch up on Nathan's podcast, but I, I sort of tend to listen to MP3 podcasts and not really on YouTube, but that's okay.
1:09:44 And if anyone's got any questions, email me at team.DataFam@Gmail.com.
1:09:50 But again, Nathan, thank you for that. That was awesome. It was good to see how all that comes together. I've sort of articulated that triangle quite a few times, but that presentation was, was awesome how you've built that and, and explained that really well.
1:10:03 Nathan Settembrini: Thanks for having me, Chris.
1:10:05 Chris McClellan: Awesome.
1:10:07 And if anyone wants to chat, now's the time, or that's the end of the meeting for today.