Team DataFam Tableau User Group 15 Sep 2026

0:01 Chris McClellan: So welcome to Team DataFam Tableau User Group meeting number 92.

0:05 Uh, this TUG works a little bit differently to others, 'cause we aim to capture all the in-person unrecorded presentations and give them a wider audience. So I don't really care if you've presented before or not presented before, or a large conference or a small conference, it doesn't matter. Everyone is welcome here, and we'll find a time slot for you.

0:24 Just get in touch with me at team.DataFam@Gmail.com and we'll talk about a time. But today we have Danni presenting why some visualizations are instantly understood.

0:37 Um, again, talking about visualizations, intuitiveness, practical examples, et cetera.

0:43 Danni's the expert on this, so I'll hand it over to her and she can take it from here.

0:48 Danni Liu: Thanks, Chris.

0:50 So thank you for the introduction, Chris. Um, I'm just gonna…

0:54 There's a lot of familiar faces on, on tonight, but I know that there's also some, um, unfamiliar faces, so I'm just gonna give a brief introduction about myself before I jump to this particular question that we're gonna answer this evening, this afternoon, this morning.

1:13 So a little bit about me. I've sort of embarked on a, as you can see, a quite a nonlinear career trajectory. I started out as a pharmacist, uh, did practice for a few years, and then I decided to move into corporate where I tried my hand out at quite a few different roles in pharma.

1:34 And somewhere along the line I got too interested in Excel.

1:38 So that eventually pulled me into analytics. And in analytics I discovered Tableau.

1:43 And Tableau gave me this creative, like, outlet that I've never really had in many of my previous role.

1:50 It's a bit like a magician's tool in the sense that it allowed me to take something like a very messy data set with, like, hidden nuggets, and then transforming that into gold.

2:02 What I love about, like, I guess what I could do with Tableau, uh, and data visualization, is to kind of see when that, you know, um, idea or something that's complicated land and someone just get, gets it. I mean, I love seeing that moment. And that's why I've, you know, kind of dedicated this, my future ti- career into sort of getting better at data visualization and information design because just the power of having impactful visual, um, is, um, is very persuasive i- in my opinion.

2:40 And it can communicate, um, I guess it can create changes.

2:47 So which is, which kind of like, I guess, that's kind of led me to start my own business, uh, called DashByte, and the… my overall objective is to help people see things more clearly.

2:59 Which ties in very nicely to the topic that we're gonna cover, why some visualizations are instantly understood and others aren't.

3:08 Now, um, we don't have, I guess, that much time to cover the depth of this, um, to answer this question.

3:16 So instead of, like, me going down into any one rabbit hole, I'm gonna keep us fairly focused. Uh, so consider this session a bit more like a tasting plate. Um, I want to give you enough of the underlying ideas to see how things kind of fit together, and also show you what you can do with those ideas.

3:37 My hope really is to leave you with some useful mental model, as well as a few practical ideas that you can take back and apply, and maybe a few things that you want to sort of take with you to explore further.

3:51 Okay?

3:54 So I'm gonna ask you a question. Um, when it comes to bad visualization, what normally comes to your mind?

4:05 Feel free to put it in the chat or unmute and then call out.

4:12 Do any of these look familiar?

4:15 So a lot of these, um, classic bad visualization, they're somehow food related.

4:22 We have overloaded pie, cluttered bars, and then the crazy spaghetti line.

4:29 Now, there's obviously a… they're pretty carb heavy bunch, but that's not the problem. There's just so much going on, right?

4:42 But I would say that most of us can probably label them as bad charts just by looking at them.

4:49 I'm gonna show you another chart, and let me know what you think of this one.

4:54 I'm going to say at first glance, this looks like a pretty decent chart. It's clean, it's readable.

5:01 There's not a lot of color.

5:04 Nothing obviously wrong.

5:07 Yeah?

5:10 We're gonna ask you a question.

5:15 Who exceeded their target?

5:18 Now take a few seconds and have a look.

5:21 Who exceeded the target?

5:26 Took a while, didn't it?

5:28 If you were anything like me, your eyes are probably moving from backwards and forward from actual target value for one person, do that for the next person, and then compare it to the previous person.

5:40 There's a lot going on in your head.

5:44 Now I'm gonna show you a different version, same data, but I want y- you to answer the same question.

5:52 Who exceeded their target?

5:55 Much easier, right?

5:57 We can tell that… We can easily see that Marcus, Ethan, Priya, Sophie exceeded their target.

6:05 And we can easily tell why, how much they exceed the target, even without looking at the, the numbers that you see on the right-hand side.

6:17 So why did the second chart feel a lot easier?

6:21 We have exactly the same data, same question, but completely different amount of mental, uh, uh, work that's required.

6:30 To answer this question, we have to talk briefly about visual perception.

6:35 Even though I studied pharmacist, I promise I'm not gonna turn this into a neuroscience lecture. I don't think I know the level of detail.

6:44 So what is visual perception?

6:49 Very simply, visual s- perception is our brain's ability to make sense of what our eyes take in.

6:56 So really when we, when we're seeing, it's not really happening in our eyes so much, it's actually happening in our brains.

7:08 Why does that matter?

7:12 It matters because our visual system isn't neutral.

7:16 It has certain tendencies on how it notice patterns, how it groups things together, and makes sense of what we see.

7:26 Now, understanding some of these tendencies help explain why certain visualization feels so much more intuitive than others.

7:36 We don't need to be an expert in visual perception to be good at vi- to become a good visual information designer.

7:46 But there are three useful ideas you need to know and keep in mind.

7:51 They are, number one, some visual features are noticed very quickly by our brain.

7:58 Number two, our attention is selective.

8:02 Number three, our working memory is limited.

8:07 Let's take a closer look at all three of these.

8:11 Starting with the first one.

8:15 What you see on the screen here are visual features that our brain can detect very, very quickly before we're even consciously aware of them.

8:26 Some of you may have heard of these being referred to as pre-attentive attributes.

8:32 All right. Position, length, color, size, shape.

8:36 These are the common things that we apply when it comes to visualization, but there's actually more.

8:46 And it's not only these features that our visual system picks up.

8:50 Our brain also automatically organizes what we see.

8:57 Here, for example, both sides contain exactly the same 18 dots.

9:04 The only thing that change between image one and image two is simply the spacing, and by changing the spacing, we're changing how those dots appear to be grouped together, right?

9:19 What I'm showing you here is a principle, a, a Gestalt principle of proximity, which states that things that are close together tends to be perceived as belonging together.

9:34 Gestalt principle came out of Gestalt, um, school of psychology back in the '20s, the 1920s, I believe.

9:42 And what they did is studied a bunch of, um, elements how our brain sort of auto-group separate visual components into meaningful whole.

9:52 This is just an example of one of the Gestalt principles.

9:58 Now, there are many that's important for us to know, but, uh, the point that I want to make h- here is that before we're consciously interpreting information, the design is already shaping what we notice, h- and how we perceive the structure.

10:16 Much like, you know, what they say about you have, like, three impr- uh, three minute, three seconds or something to, to, to make an impression on someone already. This is kinda that case as well. Your design is already being sort of noticed in this, this perceived structure before our brain is even, like, consciously digesting that information.

10:41 Now to the second idea.

10:43 This second idea is much about attention being selective.

10:48 When we look at the world, it feels like we're taking in the whole scene at once.

10:55 But that's not the case.

11:00 What it is is that our eyes are actually constantly moving, fixating from one part of the scene to the, then to another, then another.

11:10 So in reality, we're kinda taking samples of the world, and it's our brain that's kind of piecing everything together.

11:19 And where we look, these fixation points, they're not random.

11:24 It's closely linked to our attention.

11:28 And we're gonna do a quick demonstration.

11:32 Take a look at this image for a second, and where did your eyes go first?

11:41 I'm gonna… Because I can't see everyone, can I ask, um, people to unmute and then call out what they see?

11:51 Shaun Williams: I see yellow first.

11:53 Danni Liu: Yeah, yellow.

11:55 Shaun Williams: Yellow.

11:56 Danni Liu: Yeah. Anything else that's drawing your attention?

12:02 Shaun Williams: Red carpet.

12:04 Danni Liu: Red carpet, yes.

12:07 Anyone else?

12:10 Chris McClellan: And in the chat, David's saying the red carpet. And, um, Mari- sorry, Marine? I can't even say that name properly, sorry.

12:17 Mariangela: Mariangela.

12:17 Chris McClellan: Maria? Sorry.

12:19 You can say it.

12:20 Mariangela: No.

12:20 Chris McClellan: Hi.

12:23 Mariangela: Yeah, the- the blue jumper against the red.

12:26 Danni Liu: Yes. Yes. Very good. I think you've picked out all the obvious points. Um, so yes, the curtain, the yellow curtain, the women in blue. Um, maybe also the- the little monk in the front, in the foreground, versus like a lot of the things that's in the background. And the leading carpet is definitely another one.

12:46 Now I'm gonna give you a slightly… Uh, I'm gonna give you another task.

12:53 Now find the pot plant.

12:58 Chris McClellan: Near the stairs.

13:01 Danni Liu: Who was that? I feel like I need to find another, like, you know, task.

13:05 Everyone's picking out that really quickly.

13:09 Again,

13:12 I'm gonna ask everyone to kinda cast their, um, sort of, I guess, attention back a little bit and notice how after I've sort of given you a task, your attention narrowed and you- you almost like started scanning the s- scene for anything that might look like a- a pot.

13:32 And so during those moment, a lot of the other detail becomes less relevant, right?

13:39 So the point that I want to make here is that our te- our attention is really shaped by two things:

13:48 what stands out in the visual that we're looking at, and what we're actually, actively looking for. So when it comes to design, not only do we have to sort of keep, um, sort of, uh, think about the- the, um, how to draw attention externally, but also think about the user. What are the things that's going through their mind as well?

14:20 So the last idea that I want to touch on is our working memory is limited.

14:26 At any one time, we can only po- the average person, I'm not talking about those who can, who has like a photographic memory. We can only hold three, roughly three to four meaningful chunks of information.

14:40 Uh, we have like a very small, I guess, workbench, if you will, which let, um, let us hold only a few useful information.

14:49 And a lot of the visualizations that we see take up that limited space more than we realize.

14:56 You know, we saw that in the opening, um, example, the sales example that I showed, um, where I- I asked you to do the task. We're us- utilizing a lot of your sort of limited working memory to hold on numbers and to piece things together to compare things in your head.

15:15 Now, we don't want to do that. Like the real work… We want to utilize your, I guess, limited working memory to do something that's more meaningful, like interpreting the information. So when it comes to, like, good design, we want to try to reduce the unnecessary mental overhead so that we can use the limited capacity towards thinking that matters.

15:46 So now that we've kinda had a, like a quick preview of all these three ideas about visual perception, so what? What does that actually mean for how we design?

15:59 Well, these ideas give us a clue about the design itself, what the design itself needs to do, like guide attention, make relationships easier to see, and organize information in a way that supports the task.

16:16 And that translates into three questions I tend to ask when I'm designing.

16:23 What should stand out?

16:25 What patterns or relationship needs to be visible?

16:29 What should the information– How should the information be organized?

16:37 Let's go into th- each of these now.

16:42 So with the first question, what should stand out? This is really about directing attention to what matters.

16:50 Okay?

16:51 And earlier, I shared, um, the idea about pre-attentive attribute. It is a- an extremely useful tool to use to draw attentions.

17:02 Things like your color, size, orientation's one that you probably haven't seen earlier.

17:07 But they're useful tools for creating emphasis and drawing attention.

17:13 But there's an important catch.

17:16 What stands out depends on what surrounds it.

17:21 What do I mean by that?

17:24 You can see that here. So on the left-hand side, the image number one, the surrounding elements are quiet and consistent, so the different circle is very easy to notice.

17:39 Now contrast that with image two.

17:42 It's a lit- it's a little bit harder to know where to look because there's a few different variations in size and, uh, uh, different in color as well.

17:56 So the point that I want to make here is directing attention isn't always about making the important things bigger, louder, um, which is our natural inclination.

18:11 The more sophisticated way of doing it is just about making everything surrounding it quieter.

18:19 So you want to create that contrast.

18:23 Let's take a look at an example, um, in practice.

18:29 This is a service ticket trending down chart.

18:32 Now, I'll give you a few seconds to, to sort of take in the information.

18:45 Wouldn't you agree that there's a lot of things that's competing for your attention? We have the KPI cards on the right. We have the, uh, color indicator. The confusing thing about the color indicator is the arrow and the color convention, they're flipped, which is something that's causes a bit of a confusion. We have the crossbar being the quarter markers.

19:10 We have the grid line. We have the trend line. Then we have the yellow highlighted points. Oh my God, I'm tired just saying it.

19:19 So there's a lot of visual competition, right?

19:24 It's not immediately clear as a viewer or audience to know where should I look.

19:32 I'm gonna give you an improved version to look at.

19:37 This is the second rendition.

19:40 We're starting to sort of quiet down a lot of the information.

19:45 We're simplifying the supporting element, and so the KPI information's kind of folded into the main chart as well as the title.

19:55 Okay.

19:58 Let's look at the final rendition.

20:03 This is far more refined.

20:06 The trend line disappeared. The max and min is receded to the back end, but it's still there.

20:15 The main message becomes a lot clearer to see. You know, like four point six percent de-decrease year on year, point nine percent increase month on month. And you have the chart that s- is synchronized with that message.

20:32 So notice that all the information that you see in the refined version compared to the earlier version, all that information is there.

20:46 What change is actually the hierarchy of the information. The important things now have room to lead.

20:56 So the mess- the takeaway message is let the im-important elements lead, all the supporting elements recede to the back.

21:08 So once we've directed attention to what matters, the next question is whether the pattern or the relationship is easy to see.

21:16 This is where we kinda talk about chart choices and more broadly, how we kinda include the data.

21:23 Here, I'm showing you the same values in three different ways. The numbers haven't changed at all,

21:33 but notice our ability to judge the difference between those value has.

21:41 Let me get a little pen.

21:46 So we can easily tell what this, what this value is opposed to that, but it's so much harder to do that when it comes to a circle area. It is still…

22:03 It's a little bit harder as well t- if we're talking about color here, co-color grading, right?

22:16 Well, let me erase everything, erase ink on the slide.

22:19 Now, there's two statistician that published a paper, and they studied this, um, back in the e- 1980s, I think, uh, this idea of graphical perception, which is graphing all these values and how accurate we can judge the values.

22:36 And they came up with something like this. This is my rendition of it.

22:41 But, um, so some visual properties are more, uh, sort of we can judge them more easily and more accurately than others. So for example, um, you saw earlier the column where we ha- uh, the column, uh, chart, we had, um, the same scale. It was very easy for us, and we could accurately kinda compare the, um, what the value is, as opposed to, say, things like, um, the area was a bit harder to judge.

23:14 And then when it came to shading and saturation, is a little bit harder again.

23:19 So this is kinda like, um, if you want something that's more precisely represented, choose chart that's, uh, on this end and less precisely on that end.

23:30 Now, that doesn't mean that we should always choose the more precise encoding.

23:35 The right chart choice still depends on the audience and what you're trying to sort of achieve and help them understand.

23:42 So choose the visual form that best supports the purpose.

23:47 Now, I know that chart selection is challenging for a lot of people, um, so I've included this for anyone who, who is struggling or find it challenging. This is a very helpful visual vocabulary tool by Financial Time. And, um, it would allow you to sort of pick the charts.

24:09 Right. Happy to, uh, uh, talk, uh, a talk on this if or if anyone wants help on this, I can sort of walk you through as well.

24:19 Now, chart choice is only part of the story.

24:25 The other thing that, uh, we could, need to think about is how we arrange the data can actually make a pattern more easier or much harder to see.

24:37 What do I mean by that? Let's have a look at an, an example.

24:43 So we're looking at male and female salary across five pay grades. In this particular version, you can see that, uh, the gender gap, um, increases as the pay scale goes up.

25:00 I want to sort of ask you to look at the gaps. How easy is it to see the gaps in each of the pay grade?

25:09 Take note of that.

25:12 And we're gonna switch to the second version.

25:17 So right now we have female versus male. We're gonna switch to the second version.

25:23 Now we have male versus female. But notice that the gap is actually a lot easier to see in this particular version.

25:32 So if we were…

25:35 Like, would you agree?

25:36 I hope you, you agree.

25:39 Um, so I'm just gonna switch back so you can kinda see the difference.

25:44 This is first version.

25:47 This is second version.

25:52 First version, second version.

25:58 So why is this gap more obvious in the second version?

26:06 It has a lot to do with the Gestalt principle of continuity, where you see that smooth curve in the first one, that even though it's comparing the different pay grade, but the way that the data…

26:24 I mean, the actual data forms a very smooth curve. So our brain is likely to perceive it as like a unity.

26:34 This is not what we want because it's kinda dominating our, our, our thinking.

26:39 For this particular purpose, where we want to sort of, I guess I'm assuming the message is more that there is an inq-inequality in pay gap, we need to do something.

26:50 For a more impactful visual, the second version work-works better because by switching the order, we're breaking that continuity and the wholeness.

27:00 And so we're able to see each of the bay, pay, uh, grade, the male, female as a one group.

27:11 Uh,

27:14 so when we want to make a pattern obvious or a relationship obvious, don't always just think about chart, also consider how you, the data is arranged.

27:30 So to the third point. So, so far we've been looking at a lot of the individual visualization. But once we sort of bring several visuals together into a dashboard, there's another challenge, what I always find it quite challenging, is how do you sort of, um, how should you put all the information together? How do you organize it?

27:52 Um, and these are like…

27:55 So this is something that I struggled and I thought a lot about, and this is what I've come up with.

28:00 I usually then start with like a process of asking three questions to help me, um, think about how I should group these informations.

28:10 These three questions I ask myself, um, is who is using it?

28:14 What are they trying to work out?

28:16 And what do they need to keep in, in the same view?

28:24 So once I understand that, I then kinda group a lot of these related questions together and build the dashboard structure around these groupings.

28:35 Now, that may sound a little bit abstract, so to make it more concrete, uh, let's look at a demonstration dashboard that I built. Um, it's, uh, it's a fictional f- it's essentially a marketing campaign, uh, dashboard for, um, a fictional pharma company. Everything about it is fictional.

28:57 A little bit of background about this particular fictional scenario.

29:01 The product that we have is called Ricardus. Um, it's used in chronic kidney dis-disease and type two diabetes. It's been on the market for around twelve months, and the uptake has flattened.

29:16 And so the campaign objective is to rebuild that momentum, and the, the team has identified three target healt-healthcare professional groups, being the nephrologist, the cardiologist, and your general practitioner across several, uh, channels.

29:39 Now, when I design any dashboard, I always have a couple, one to two, maximum three user groups in mind, and they represent each of these user groups. For this particular dashboard, I had two user group in mind.

29:57 Lar…

29:58 The brand manager and the sales group, and we have Lara representing the brand manager and David representing the sales manager.

30:07 Now, their roles are different, and obviously then their responsibility will be different and the care factor of different things will be different, right?

30:15 So for Lara, who's a brand manager, her priorities is more around the campaign performance.

30:22 Whereas David, given that he's a sales manager, his care factor is more around how well did the team execute the campaign.

30:33 So instead of, I've seen some dashboard in my li- life around super long dashboards, uh, that cover, tries to cover everything.

30:42 I've not done that. I don't like it. So what I've done is essentially organize the dashboard into three views.

30:50 Here, the way that I've organized it is into, uh, a view called campaign overview, channel performance and efficiency, and field execution and coverage.

31:02 And we're gonna look at each of these views, and I'm gonna sort of point out some of the design elements to it.

31:10 So the first view.

31:13 Oops.

31:18 We're gonna give you, like, a few seconds, maybe five seconds to sort of take in the layout.

31:26 Now, this campaign overview, uh, view of this dashboard set really answers one question: How is the campaign tracking overall?

31:40 So instead of treating, like, a collection of charts or groups of, like, channels, I've kind of grouped the information around the questions that someone looking at this is likely to have and sequenced them in that order they're likely to see, uh, likely to c- uh, ask.

31:56 So we start with, "Why are we running this campaign, and where are we up to?

32:03 What are we trying to achieve?

32:07 How are we executing, and how is this translating in the market?"

32:13 As you can see, it's, it's quite logical, so it provides, like, a clear path.

32:22 The other thing that I try to do with all my dashboard is to have this two level of reading, macro reading and micro reading.

32:32 What do I mean by that?

32:34 So with the macro reading, it's more the overarching big idea so that people can kinda see what the pattern, the signal, the status is. And the micro, uh, reading kinda gives you the detail, um, that someone might want to res- uh, inspect or explain what they're seeing.

32:55 Now, often good visualization gets the right balance of macro and micro readings that dance right.

33:04 And, um, the way I like to kind of think about the macro and micro idea is that a view has a dedicated purpose. So if we're playing video games, it's like the main quest. My main quest, I have a particular goal.

33:19 But occasionally if I, you know, venture through my main quest, I would do like a little side quest because something sparked a, a question in my head.

33:28 Um, like something, there's a curiosity that's kinda triggered during my main quest, and I want to sort of satisfy that side quest.

33:37 So side quest is where you kinda give them that micro information.

33:44 The example I have on the screen is when, say, someone looks at execution, we've completed 11 events.

33:52 That is likely to trigger in someone's head, "Oh, what are these 11 events?" So I've, I've created like a little hover that actually shows the 11 event so that they, it would satisfy their curiosity so that they can go back to their main purpose, which is to understand the overall performance of the dash- uh, of their campaign dash…

34:16 campaign execution, sorry.

34:20 Now, to the second view, which is channel performance and efficiency.

34:25 There's two main questions around this particular one, so I'm gonna give you another five seconds to kinda take in the layout.

34:34 So I've organized the view around two related questions.

34:40 On the left, you have, "How is each channel performing?" On the right, it's about how each territory are performing, um, with regards to each of the channels as well as each of the specialty groups that they're targeting.

35:00 Now, just quickly, I'm gonna point out, notice I've also created that macro read and the micro read. So on top, the channel cards that give you the high level read, and the bottom gives you the detail of any of, um, if any of these channel aren't performing or if there's anything that creates curiosity.

35:24 But the main thing that I actually want to sort of call out here is around the territory comparison on the right.

35:30 These are called small multiples or trellis charts, so they have exactly the same axis, and each territory uses the same frame. So we're able to sort of compare things quite quick- quickly just by looking at, just by looking at all the information.

35:48 So we're not asking the, I guess, the viewer to sort of remember things about previous territory. Everything is kinda laid out in front of them. This is to enable quick patterns, uh, spotting, um, as to, you know, which one is driving good performance or which one's not underperforming. And when they see the detail, they can then go to the filter on the top left to look at the territory which has a lot more detail.

36:20 Now, like, the driver behind this is I want to sort of limit, um, I don't want to create, uh, maximize, um, the limited, like, working memory that someone has, and also sort of surface all the patterns.

36:41 The last view we have is the fi- field execution and cov- coverage.

36:48 Earlier I introduced Lara and David. They're the representative of the two persona groups that I have when I'm building this dashboard.

36:56 But know that that doesn't mean every view has to serve them equally.

37:02 I would say the earlier view was more for Lara. This particular view is more for David.

37:10 And knowing that this view is more for David, I would think about the questions that he would need to sort of, um, answer.

37:17 Like what needs attention of all the, say, the, the objective of the campaign?

37:23 Like, which territory isn't performing, and where are the specifics?

37:30 So having those, like, questions in your mind, it really guides out or guides your layout and the shape of your layout.

37:42 So these, the points that you see, um, highlighting the, the, the questions that David has, right? Which objective needs attention? So I've…

37:52 So that he doesn't have to remember what is the objective, it's there.

37:56 It also gives him the high level number, KPIs.

38:00 Is he hitting that?

38:02 Where, where, which territory are falling behind and which particular channel of that territory is falling behind? And then number three, he could see the details behind the aggregated number for that territory and for that channel.

38:19 So the way that I've designed this is I'm not making David scan all the detail in the dashboard. I'm actually kinda guiding him through the questions that he's likely to ask so that he can get to his answer a lot quicker.

38:34 Now, when you structure the, the flow, or align that to someone's, the questions that's, uh, going to, the questions that the person have and the, the order of the question, it feels a lot more natural for someone who's using the dashboard.

38:55 And that brings us back to the, our question: Why are some visualization instantly understood and others aren't?

39:04 A big part of it is whether the design works with the way we naturally see and process information, and when it does, things tend to feel more intuitive and at ease.

39:16 When it doesn't, we create a lot of friction.

39:21 And those three questions that we visit through this session: What should the viewer notice?

39:27 What do they need to see or understand?

39:30 Is the information organized in a way that helps them do that?

39:36 That's a simple way for us to put what we know about visual perception into practice.

39:43 So I'd like to close this presentation with one question to you all.

39:50 How much work are you asking your user to do that the design could have done for them?

39:56 Thank you.

39:58 Chris McClellan: That's always the hardest challenge, right?

40:00 Danni Liu: Yeah. It's not easy.

40:01 Chris McClellan: Designing is so-

40:03 Danni Liu: But I would say having a lot of empathy for the people that you're designing the dashboard or visual for really actually helps. So if the Mo- the better you can actually come from their angle, um, and view it from their perspective, it really helps you craft what you need to do.

40:23 Chris McClellan: Sure.

40:26 Danni Liu: Cool. Does anyone have any questions?

40:30 Shaun, you have your hand up.

40:31 Shaun Williams: Hey, Danni. Thanks very much. It was fantastic presentation.

40:35 Um, a lot of things to kind of take away and think about. I kind of had a, a question, um, uh, on intuitiveness-… versus, um, maybe how much the user was familiar with the data, a- as in, um, y- you know, how much familiarity would you kind of, uh, associate with someone being able to interpret a chart f- you know, versus, um, if they're seeing it for, say, maybe the first time, how much the design plays, plays kind of an impact there?

41:12 Danni Liu: I think it's then coming down to how well you understand who you're designing it for. And so when you have like a persona in mind, you also, um, not only are you assessing, I guess, their roles and responsibility, but you're also factoring in, I guess, how savvy they are in terms of like data technology, and you'll factor that into your design as well.

41:34 So if they're not familiar with the data, perhaps you could have a bit more like pop-ups or interactions to kind of guide them and help them understand the data. So that comes in, um… I always, um, when it comes to like… I, I guess 'cause when we're creating stuff, we're creating for people, right? So the better you understand your user, the more you can factor those things in.

41:55 I hope I've answered your question, Shaun.

41:58 Shaun Williams: Yeah, no, it's…

42:00 Danni Liu: Does anyone else have any questions?

42:03 Chris McClellan: So Zilla was saying, "Love the examples, and shows that these data visualizations are not just data, but a lot of art."

42:11 Danni Liu: Oh, yes. Yes.

42:13 Yeah. I think that's why I like it, um, like the data visualization space. There's an analytical piece, but there's also the very subjective piece around like the art, right? It's a, it's a fine balance.

42:25 Um, uh, yeah, that's why I find it quite interesting. And I guess my previous role didn't give me that creative license. That's why I like it.

42:35 David Nguyen: Um, thank you for, for your presentation, uh, Danni. Um, uh, so I did have a quick question.

42:40 Danni Liu: Yeah.

42:40 David Nguyen: So I really like the idea o- of how you organize, you know, um, the dashboard to the different, uh, session that answer like different, uh, questions to different audience.

42:50 So a very quick question.

42:52 Danni Liu: Yeah

42:52 David Nguyen: Uh, let's say if a stakeholder, like, they ask too many questions, five or 10 different questions, so how, how, how do you balance that? Or how do you, like, push back a request, like, you know? Um, a- a- any limitation, any restriction on how many, uh, uh, question is put, you know, on the dashboard? Something like that.

43:09 Danni Liu: Ah.

43:10 David Nguyen: Yeah.

43:10 Danni Liu: Yes. Are you saying how do you meet all users' needs, right? They ask, they ask-

43:14 David Nguyen: Yeah. Yeah, yeah

43:15 Danni Liu: … ask. Okay.

43:16 David Nguyen: Yeah.

43:17 Danni Liu: So, uh, it is part of, I guess, um, the dashboard development journey.

43:22 David Nguyen: Yeah.

43:22 Danni Liu: And, um, this is adopting from design thinking.

43:26 David Nguyen: Mm-hmm.

43:26 Danni Liu: During the interview process, you, it, what you want to sort of explore is the reason for the request or the feature.

43:33 David Nguyen: Hmm. Yeah.

43:35 Danni Liu: Yeah? Um, because-

43:36 David Nguyen: Yeah

43:36 Danni Liu: … they have certain request and feature, but if you can actually understand the underlying re-

43:41 David Nguyen: Agreed

43:41 Danni Liu: … reason for the request, it'll help. And a lot of time, even though they have multiple questions, right?

43:47 David Nguyen: Mm. Yeah.

43:47 Danni Liu: But there's themes to that questions. So collect all the questions and the reasons-

43:52 David Nguyen: Yeah

43:52 Danni Liu: … behind those questions, organize them into a theme. Um, and then-

43:56 David Nguyen: Mm-hmm

43:56 Danni Liu: … craft your, your dashboard. Does that make… Does that help?

44:00 David Nguyen: Yeah, yeah. Yeah, yeah.

44:01 Danni Liu: Yeah, yeah.

44:02 David Nguyen: Thank you.

44:02 Danni Liu: So it, it's kind of like the interview process, but-

44:06 David Nguyen: It's kind of interview process, yeah

44:07 Danni Liu: … my suggestion… Yeah.

44:07 David Nguyen: Yeah.

44:07 Danni Liu: And my suggestion for the interview, you want to do individual interview, but you also want to do-

44:12 David Nguyen: Mm-hmm

44:12 Danni Liu: … group interviews.

44:13 David Nguyen: The group interview.

44:14 Danni Liu: And don't just listen to what they say.

44:16 David Nguyen: Oh, mm-hmm.

44:16 Danni Liu: Also study some of the artifacts of the documents and, and like their workaround, because they're-

44:21 David Nguyen: Yeah

44:21 Danni Liu: … very insightful.

44:23 David Nguyen: Yeah, yeah.

44:24 Danni Liu: Yeah.

44:24 David Nguyen: That's cool. Thank you, Danni.

44:27 Danni Liu: Any other question?

44:31 Shaun Williams: Maybe, Danni, I can ask one more if, if I'm allowed.

44:35 Danni Liu: Yeah.

44:35 Shaun Williams: Yeah, I, I was like really, uh, kind of blown away when you were changing the chart types or how much information you had. And although I myself don't use dashboards that much, we have a lot of, um, kind of presentations of data.

44:50 Danni Liu: Yeah.

44:50 Shaun Williams: Would you kind of recommend like one or two messages per, per chart, you know, or breaking it up if, if you have, you know, areas to just display, or, or three, three kind of points per chart? Is that too much? You know-

45:06 Danni Liu: Yeah

45:06 Shaun Williams: … paring it more to one also?

45:08 Danni Liu: Yeah. So generally when I, when I create dashboard views or presentat- or even just presentations, right, um, I try to have like a, a core message. What is the core message? There u- there's usually one core message, and then the rest are ag- again, supporting that core message.

45:31 And if there's like multiple like key messages, you're kind of competing w- what you want the, the, I guess the, the audience to take away. And this is where I would kind of suggest you to sort of break it into the next slide. But when it comes to dashboard, what… Like, the way that you want to break up this core message, 'cause you only have finite space, right?

45:52 Shaun Williams: Mm-hmm.

45:52 Danni Liu: You have to make a judgment call in terms of do you want everything in the same view? If you want everything in the same view, um, for whatever reasons, like, you know, if you, if you need someone to recall something or compare something, have it in the same view, but you want to create clear zone for them.

46:06 Shaun Williams: Mm-hmm.

46:08 Danni Liu: Right? This is… You, you want to kind of, against the hierarchy aspect, the clear, um… I mean, I can talk, also talk about layout, because there's certain ways to structure a layout where, hey, people know this is look at me p- uh, piece, and this is all the, like the supporting element. So when you create, like sub, um… When you have that clear hierarchy, it also sends your message, uh, through quite quickly, uh, easily too.

46:36 So, but if it's like-

46:37 Shaun Williams: Thank you. That- that's enough

46:37 Danni Liu: … slide presentation, I know, I know you don't develop dashboards, Shaun. But, um, again, uh, one key message per slide, and then supporting points to that key message. And then if it i- if it has to be, like, more than one key message, because you're, you're kind of… It, it also depends on whether you're doing it like a presentation-led, like a, a speaker-led presentation versus like a info pack, right?

47:02 Shaun Williams: Mm-hmm.

47:02 Danni Liu: Because then you can, if it's an info pack, you can pack a little bit more. But again, I would like to always try to keep, keep it contained to two to three.

47:11 No more than three.

47:16 Chris McClellan: Yeah, if you overload it too much, then it gets too complicated and too difficult to understand.

47:21 Danni Liu: Yeah, yeah.

47:23 But three is my magic number. It's like one to three.

47:26 Chris McClellan: Okay.

47:27 Danni Liu: Even though I like pairs two.

47:30 Chris McClellan: Awesome. I think that's the end of the questions.

47:33 Any more?

47:35 No? Okay. We'll move on.

47:38 So there's all Danni's contact details, so LinkedIn and her two websites, which is always good.

47:44 Danni Liu: One is more the business, um, and then the other site, the website, um, it is more like personal website. Just kind of blog my designs and stuff.

47:53 Chris McClellan: No, no, that's cool.

47:55 And survey time. So if people haven't done the survey yet, there's been a few people that maybe have not done that, time to grab your phones, scan the QR code. The link's been in the chat for a while. Um, I'm not gonna start the countdown. I'm just gonna talk slowly while we're doing this.

48:12 Um, so here we go.

48:15 Um, so again, as Danni said at the start, what are the standard questions? We always start off with the standard questions. Where do you live? S- or state and country usually. So we've got an interesting mix. Australia, both New South Wales and Queensland, Dublin in Ireland, and Tallinn in Estonia, 'cause Külli's from Estonia.

48:39 Danni Liu: Oh.

48:39 Chris McClellan: Very good.

48:39 Danni Liu: I've heard Estonia's lovely. I would want to visit one day.

48:43 Chris McClellan: And the next question is what industry do you work in? So it's sort of across the board, right? Which is what the, the good thing is for Tableau. So pharmaceutical, healthcare, consulting, IT, media agency, pharmaceuticals again.

48:58 Uh, it's all there.

48:59 And our third standard question is how many years have you, have you been using Tableau? And I always joke at this graph, it's always different. So we've got people that have, haven't been using Tableau for very long at all, and people who've been using Tableau for a very long time. And you just migrate through that graph as you use Tableau more, right?

49:18 Everyone started off at zero, and we've just got experience over the years as well. Just depends on how long you've been using it for.

49:26 But the quiz questions for today are: when placing a legend, where should it be positioned for optimal readability? Um, I think the correct answer is near the related chart.

49:40 And I think that that graph just moved, so someone- someone's actually doing this quiz at the moment.

49:45 Danni Liu: I'm gonna ask a, a, a, a, a more pointed question.

49:49 Chris McClellan: Sure.

49:49 Danni Liu: When you say, um, near the related chart, where about near?

49:55 The top left, top, top right, bottom left, bottom right? They could all be near, but where?

50:02 Chris McClellan: Sure. But, but those are those other two que- two answers, right? There's like the bottom right corner-

50:06 Danni Liu: Good

50:06 Chris McClellan: … or the bottom left corner of the dashboard. But yeah-

50:09 Danni Liu: Oh

50:09 Chris McClellan: … just near the chart. So close enough so that it's, it can relate to the whole thing, right?

50:14 Danni Liu: Oh.

50:15 Chris McClellan: If you're using legends. Sometimes people say that it's best to not use a legend, but then you've gotta have some-

50:21 Danni Liu: Oh, yeah. You've still need it. Yeah

50:21 Chris McClellan: … so even on that scatter graph that you had is like, what, what do those symbols mean?

50:26 Or are those symbols just for that industry or for those users, do they understand what all those symbols mean straight away?

50:33 Uh, the next question was, in a dashboard that uses a parameter-driven calculated field to switch measures, ah, which best practice ensures the tool tip displays the correct measured name?

50:45 And again, it's moving a little bit. Someone's still answering.

50:48 Um, I thought, I thought the correct answer was create a case-based calculated field that returns the measured name and references it in the tool tip, 'cause that's what I would do all the time. I don't know.

51:00 Uh, a lot of people said insert the parameter value directly into the tool tip text.

51:06 Okay.

51:07 Um, yes, I can see why that would be the right answer, but we've chosen one, so that's, that's sort of bad news.

51:14 Um, the next question…

51:17 Okay, cool.

51:18 I can see a glitch coming up, so I know everyone else is seeing the glitch as well. Um, which of the following is considered a poor practice when designing a Tableau dashboard? Um, adding unnecessary 3D effects.

51:30 Everyone knows that one, right?

51:32 And the next question is, uh, which of the following is considered a poor practice?

51:37 I've obviously duplicated the questions and I didn't even realize it, and everyone got-… that same question right again, which is even better.

51:46 And in a bar chart, which axis should usually start at zero to avoid misleading bar lengths? Definitely the y-axis. Both axes. I don't know if we would have a both axes starting at zero in a bar graph, but I'm not sure.

52:03 But that's good.

52:06 And more for Danni's questions: which part of today's sessions are most useful?

52:12 So the three practical design questions, the dashboard case study, and the before and after visual examples.

52:19 And if she presented this session again, what would make it stronger?

52:23 Um, just everything, right?

52:26 Because most of those things are-… people have wrote at the bottom. And people can't see the other answers, so it's, um, interesting that it came out like that. So more practical examples, more Tableau demonstrations, and no major changes. Interesting. And, uh, what was the last question? What would you like to learn more about when it comes to dashboard design?

52:43 For example, uh, using color effectively, choosing the right chart, et cetera, and there's a whole heap of things there as well. I'm not gonna read all those out, but you can go back and see those in the recording as well.

52:56 And back to our standard questions is which topics would, would you like to see presented at a future TUG meeting? So some data prep tips and tricks or a case study. Tableau AI.

53:07 Um, yes, we've had some MCP presentations, but we need to do some AI as well.

53:13 And who would you like to see present at a future TUG meeting?

53:17 I can always get Mark, but, uh, yes, we'll see what happens. Um, Shreya, uh, Hero, et cetera.

53:25 So the leaderboard based on the quiz questions, so that's some of those tonight, were Shilpesh-

53:32 Danni Liu: Shilpesh

53:32 Chris McClellan: … is first with 41 seconds.

53:35 Danni Liu: Ooh.

53:35 Chris McClellan: David is second with two minutes. Or what, what does this mean?

53:40 That's interesting. Two minutes and then f- oh, okay. I see. The points drop down as well.

53:45 So Shilpesh-

53:46 David Nguyen: I stopped for a while.

53:48 Chris McClellan: Sorry?

53:49 David Nguyen: I, I, I was doing the question and then I stop for a while.

53:53 Chris McClellan: Other two minutes?

53:54 David Nguyen: Yeah, yeah.

53:58 Chris McClellan: And, and why am I not there? Because, um, I wrote the questions, so it's really hard for me to actually answer the questions slowly because I know what-… all the answers are anyway because I've gotta build the quiz.

54:10 But anyway, again, the points don't matter. You would've learned something while you're doing that anyway.

54:14 So which is why we actually do that like that, is that anyone can actually submit questions for that quiz. Any TUG can use that quiz format as well. And if you've got questions you wanna submit, then you can use the QR code, use that link, um, talk to me at any time and we can add your questions in as well, as Danni did tonight with hers.

54:37 But talking about Mark Rekker, let's go back to talking about Matthew Miller. Both Matthew and Mark have got a competition as to who can appear in the most number of TUGs over the next year. I don't know when it really started. But, um, but Matthew is on the next TUG meeting. The next TUG meeting's on the 29th of September.

54:56 That means that we don't have anything next week. So someone might drop in the next week. But at this stage, Matthew Miller is presenting Data Demos and the Art of Storytelling in the Age of AI at those times.

55:11 Danni Liu: Oh, yeah.

55:11 Chris McClellan: So, um, I'm sorry, Külli, but it's a midnight time for you.

55:16 That may, that may mean that you wanna watch the recording. That's okay. It's gonna be an early start for people in Australia, and it's sort of just the right time for the US, 'cause that's where Matthew lives as well. Um, but you can scan the QR code, uh, to attend that, that meeting live, ask questions, see Matthew's presentation. Obviously, if you can't attend the meeting, everything that we do, and a whole heap of other, other online TUGs, and a whole heap of other TUGs as well, um, anything that, any TUG that records something is on YouTube at Team DataFam, and there's an MP3 podcast at Spotify and Apple as well.

55:51 But on that YouTube channel you can see the full recordings, the shorts, whole heap of details.

55:56 And if you've got any questions, email me at team.DataFam@Gmail.com.

56:02 Uh, and this recording will be up in a few days as well.

56:05 But thanks, Danni, for that. It was great to see that presentation and what you were talking about, just our principles, dashboard layouts, et cetera. Pretty awesome to see.

56:15 And if anyone wants to talk, now's the time.

56:17 Unmute, chat away, or if not, that's the end of the meeting for today.

56:22 Danni Liu: That's a wrap. I'm gonna have dinner.

56:25 Chris McClellan: No worries.

56:28 Danni Liu: Thank you for tuning in, everyone.

56:31 Chris McClellan: Okay. Thanks a lot.

56:32 David Nguyen: Thanks for having us, Danni. Thanks.

56:33 Danni Liu: Bye.

56:33 Chris McClellan: Okay. See you. Bye.