Greater Vancouver Tableau User Group 23 Sep 2026

0:01 Iris Sun: So again, welcome everyone for our first event. I'm so excited, and the great news to share with you, as of today, we already have 43 members. Thank you again for your support. The Greater Vancouver, uh, Tableau, uh, Tableau User Group is rooted in Vancouver, but everyone, uh, is welcome, whether you are joining us from Vancouver, City of Vancouver, uh, Greater Vancouver area, or anywhere, you know, outside of Canada, very welcome.

0:30 And I'm looking forward to, uh, grow this ta- uh, the tab- together with you.

0:37 Uh, we have over 50 people registered for today's event. If you are not member yet, please do consider joining us, uh, uh, uh, as a member as well, and then you will be receiving the emails on the upcoming, uh, events and the other information as well.

0:56 So before we get into our today's topics, and please allow us a few minutes to introduce ourselves. First, we have Den- Professor Dennis Kao. I believe many of you, if not all of you, know Dennis very well. And Dennis is, uh, the, uh, social worker professor at, uh, Carleton University based in Ontario, and you can see on this screenshot, uh, his, uh, three-time vision- Tableau visionary and, uh, four-time Tableau Ambassador.

1:27 Actually, that number four you see right now gonna become five very soon, in a couple of days, uh, with the new badges for 2026 Tableau Ambassador.

1:36 And he has nine Viz of the Day, such a high bar for me to reach. So, um, Dennis is my, my mentor, so, so I go ask him questions all the time. He's my mentor. Thank you, Dennis.

1:50 And his Tableau profile actually is my go-to. I try to not bug Dennis that often, so his Tableau profile is my go-to place for all the creative ideas, uh, all the, uh, beautiful color frames and of course, the Tableau map, uh, and the tips and the tricks on the Tableau mapping. So Dennis, do you wanna share a few me- a few words here?

2:15 Dennis Kao: Hello everybody. Thank you for joining. Uh, this is exciting. I'm based in Ottawa, not, um, but I wanna be in Vancouver, so I guess the next best thing is being a co-host of this Tableau group.

2:28 Um, Vancouver's one of my favorite cities in the world, so, um, I try to go every now and then. But, um, I'm based in Ottawa, uh, teaching social work and happy to be here and, and thank you for joining us.

2:42 Iris Sun: Thanks, Dennis. Yeah, always welcome to come to Vancouver whenever we have, uh, any in-person events or not. Okay, thanks. Then you have me. I'm Iris. I'm, I'm based in Vancouver.

2:53 I originally came from, uh, uh, China.

2:56 Um, my work has Mo- mostly focused on the community-based research and the consulting services, uh, in the not-for-profit sector. I help agencies to identify their service needs, service gaps, strengths and weakness. Um, I started to use Tableau since 2020, and it quickly became my favorite tool.

3:17 Um, so throughout my Tableau learning journey, I receive all the support and the inspirations I need from the Tableau data fam, so I feel, uh, I need to give it back. So that's the reason last year I became a Tableau Ambassador, and I'm so happy to have another year to continue serve this community.

3:38 Um, what else?

3:41 Yeah. So but also, I also joined the Tableau User Group leader community, and again, I want, just want to give back to this community, and I want to grow together with, uh, with you all together.

3:56 So that's about me. I see lots of familiar faces and names. Hello, Chimdi. Hello, Angel. Lots of face. Thank you very much for your support.

4:04 And I'm so delighted to introduce, uh, Jessica Bosch as well.

4:09 And, hmm, Jessica wear two hats today, and she is one of our co-hosts and also being our community spotlight guest. So you're gonna get to know Jessica more in just a few minutes. But for now, I just want to share, a very briefly share, uh, a little information about Jessica.

4:28 And yeah, Jessica is one of the most creative, uh, data scientists I have ever met. So I got to know Jessica from, uh, uh, from our volunteer work. We both are the committee member for the Data Changemakers, one of the, uh, now organiz- not-for-profit organization based in Toronto.

4:50 Even though we get to know each other for less than one year, I'm have been so deeply impressed and also ins- inspired by the scope and the depth of Jessica's skill in terms of the data analysis, visualization, and storytelling. I'm so excited to team up with Jes- Jessica and grow the, uh, Greater Vancouver tech together. Thanks, Jessica.

5:15 Yeah. And, uh- Thank you.

5:17 Yeah. You, you gonna, you, you gonna know Jessica in just a few minutes, so Jessica know…

5:23 Yeah. So you will have a chance to tell a little bit more about you when we, uh, proceed to the community spotlight. Thanks, Jessica.

5:32 And for today's agenda, I believe you all know the agenda very well with all the emails you have already received. I just w- so we're so delighted to have Jessica, to have Kirk, and to have, uh, Kevin, uh, with us today. And we will start with the community spotlight, and then we will move to the composable data sources. And then we will, you know, uh, Kevin will share all the secret of the Tableau dashboard design with us.

6:00 So and then we will have time for Q&A.

6:04 So time for our community spotlight with Jessica Bosch.

6:09 Uh, uh, Dennis, back to you.

6:11 Dennis Kao: Sure.

6:13 So I first met Jessica also in, I think, when viz for Social Good was still underway. Um, did a few projects and, um, yeah. Um, I think she's newly moved back to Vancouver. Is, is that correct?

6:29 Jessica Bosch: Yeah, I moved back four months ago. Yeah.

6:32 Dennis Kao: Okay. Yeah.

6:34 All right. So jealous. Okay.

6:36 Um, so we just have, uh, some, some of… some quick questions, uh, five questions just to get to know Jessica a little bit more. So could you introduce yourself briefly, and then tell us what you enjoy most about being part of the Tableau community?

6:53 Jessica Bosch: Oh, yeah. So yeah, I'm, I'm Jess, so Jessica. Can call me, yeah, however you like. I'm… Right now I just started a new job actually, um, in September. Now I'm going back to actually to my roots in, in mathematics, and doing, yeah, research on computational science, which is fun. Um, but yeah, I'm still, like, loving the, the other part, um, doing, uh, data analysis, data visualization, working with, uh, also Tableau, which I like, and this is…

7:24 yeah. These two things are, yeah, what I enjoy, uh, most besides, uh, being out in the nature.

7:32 And yeah, so, um, what I enjoy most about the Tableau community, I think it's, um, like a, um, a place that is very welcoming, very open. And it feels like somewhat like a, like a, a, a big family, so you can reach out to people. You can, um, be part of, uh, community projects, and the people, they reply, and they recognize you. And it's like a just a, yeah, a very big, uh, community with a lots of resources and very welcoming.

8:08 I think that's what I would say about it.

8:11 Dennis Kao: Great.

8:12 Uh, what's one Tableau feature you use all the time or couldn't live without?

8:18 Jessica Bosch: Um, um, yeah. There are several features, yeah, I use all the time. Maybe, hmm, maybe the one I pick is the, the parameter action.

8:30 Um, you can do, yeah, several things with it. Like, you can, um, for instance, use it in a, in a, like, a dashboard in, in Tableau, uh, to have different views. For instance, to switch parameters to different views. You can use it to, um, yeah, choose different, uh, filters or aggregation levels. Um, yeah, it's something I always use in one or another way.

8:55 And, and yeah. Again, as I mentioned earlier, like, uh, when I sometimes, um, try to figure out a certain thing, how to do a certain thing, and if you search for it, and there's… Often you find, like, a, um, a blog or a how to do something, um, with the help of other people that did it before. So then y- it's always a way to learn how to do use certain features as well.

9:20 Dennis Kao: Great.

9:21 Uh, what's your favorite Tableau visualization, if you have one?

9:27 Um, and what made it meaningful to you?

9:30 Jessica Bosch: Oh, oh, oh.

9:32 There's-

9:33 Dennis Kao: Or maybe if you have a few, you can also talk about them.

9:35 Jessica Bosch: Uh.

9:37 Yeah, I, I would say, hmm, there are several. They're all so inspiring. I mean, like, maybe I can just mention maybe one that I saw very recently.

9:49 Um, let's see. Uh, I, I'm seeing all of them. And, and, and, like, on LinkedIn, a lot of times people are sharing their work or from community projects. And one I really liked was for… based on our recent community project, um, with, um, uh, Data Changemakers. We had, like, a project, um, that was also, like, doing, um, helping out a nonprofit organization.

10:17 Um, so it was a project for, it's what's called Noise Solution, and it was, like, a company that, yeah, um, yeah, basically helps, um, people, um, who are, yeah, in, in need of help, and to, to get, like, some, some support, mental support, but also physical support. And that was a, a dashboard, uh, a couple of dashboards s- submissions made for that project, and it showed, like, the impact of the Noise Solutions project on the wellbeing.

10:49 And it showed it, like, um… In Tableau, there were a couple of projects submitted that show, like, very nicely, um, not just the pure data, but how it all connects into a, a, a, an impact story. And you can use Tableau to-

11:06 Dennis Kao: Wow

11:06 Jessica Bosch: … design it very nicely, but also use the features that Tableau has to, um, present it in an impactful way.

11:13 Um.

11:14 Dennis Kao: Yeah, it's a kind of a hard question, but the right answer would've been one of mine. No, I'm just kidding. All right.

11:20 Jessica Bosch: Uh-huh, yeah.

11:20 Dennis Kao: What advice w- uh, what advice would you give someone who is newer to Tableau and wants to become more involved in the community?

11:27 Jessica Bosch: Right. Uh, I would say, yeah, two things. That's at least what I did as well when I joined Tableau a, a couple of years ago or started learning about it. Um, so one thing is, like, um, when I tried to figure out how the, um, Tableau, um, those complex calculations work, like LODs for example, what I always did, like, I had, like, um, kind of two worksheets.

11:52 Like one worksheet that does those complica- uh, complex calculations intuitively, where I'm 100% sure, like, the calculations do the right thing. And then have another, like, worksheet which, where I'm actually using, for example, L- LODs, and then compare that I'm willing making sure that my, the use of LODs, for example, really are the right one.

12:15 And I get to understand, uh, how they work, um, also at the same time.

12:20 And the other thing what I really liked, like, um, I get, like, inspired by people that do very arty but also very impactful, um, dashboards with Tableau, so I joined, um, the community projects. There's like a page on Tableau with community projects. There are so many different, um, types of communities you can join for free, and just, yeah, play around and showcase your work to others and get feedback from others as well.

12:49 That's another way, I would say.

12:52 Dennis Kao: Great. Uh, one last question is, what's one Vancouver place you'd recommend to someone visiting for the first time?

13:02 Jessica Bosch: Oh, yeah. That would be my favorite bra- place, Stanley Park.

13:06 Dennis Kao: Mm.

13:07 Jessica Bosch: Um, if you… It's like, it has like a little, a, a big, uh, forest inside with different pathways, and on the outside you can have a walk along the seawall, and it's about 10K, the outside walk, and I, I just love it. I do it every once in a while or run through it.

13:27 Yeah, definitely.

13:29 Dennis Kao: Great.

13:30 All right. Thank you so much, Jessica.

13:33 Jessica Bosch: Thank you.

13:37 Dennis Kao: All right.

13:38 So next we're gonna move to our next, our first presentation. Um, I'm just… You can read everything, uh, about Kirk, and I'm sure he'll mention a couple things as well.

13:47 Uh, if you don't know Kirk Munroe, uh, he's currently a chief analix- analytics officer and founding partner with Paint With Data.

13:56 Um, a con- uh, um, kind of a, a boutique, I wouldn't say boutique, data, data, uh, firm.

14:04 Um, he's a first time Tableau Visionary this year, so congrats, Kirk. Um, and I, I know, I don't think I'm exaggerating, but I think he's kind of the go-to person when it comes to data modeling. I, I don't know, he wrote the book, uh, Data Modeling with Tableau. He literally wrote the book.

14:25 But if you, um, you- you'll… And what you'll notice, you know, when he starts to present, and if you haven't caught, uh, some of his, uh, some of his, uh, presentations, he just did a master class with just him. He's wrote several articles that are featured in the Flerlage Twins blog. I think he has a gift of making data modeling seem very accessible.

14:49 And, and so i- and hope- hopefully, you know, that, that, I'm sure that will come through in his presentation. And he lives in Halifax, which is one of my fa- uh, wh- uh, one- another one of my favorite places here in Canada. So welcome, Kirk, and thank you for presenting.

15:06 Kirk Munroe: Thanks, Dennis, and thanks for the intro and for everyone having me today. Um, I won't spend… I don't think I need to introduce myself anymore. Dennis set me up now to live up to that. So, um, let's, uh, um… What I wanted to talk to you about was, there's a new feature in Tableau called, um, composable data sources, which came out this year in the product.

15:31 And like Dennis says, I spend a lot of time thinking about, uh, data modeling in Tableau, and that was, we get more and more AI driven. I think it's gonna get a lot more important, but that'd be, that'd be another presentation. But for now, um, in 20 minutes, I really wanted to get specifically into why, um, you shouldn't have to use Tableau Blends anymore.

15:52 Um, so, uh, basically just wanna start, go through this presentation quickly so we can get to demo. You know, so who likes blending?

16:01 Hopefully no one, it's rhetorical. Um, you know, how did we get here? So how did published data sources in Tableau evolve basically? Um, and then why are composable data sources such a big deal? And then really jump into code. Um, so first off, um, I wanna, you know, acknowledge this. I, before, um, Will Perkins and I, Will's a six time Tableau Ambassador at this point, and I did a presentation on composable data sources at Tableau Conference this year.

16:30 And beforehand, these LinkedIn polls used to do much better for me, but, um, uh, this one, you know, c- certainly not scientific. Um, but I just was curious, the, the big three reasons I think that composable data sources are a big deal is that now you can use Tableau Prep inside relationships. Um, that's a really, really big deal. That means you can leave your data at its, uh, natural, uh, grain or level of detail.

16:56 Um, and you can also do all the kind of ETL type things you can do in Prep, um, and, and still be able to use it inside relationships. The second's this one we're gonna talk about today, which is no more blends, and then just in general governance and reusability. Um, it's interesting that they came out to basically a third, a third, a third. So, um, so just to, like how did we get here that we needed composable data sources, and then specifically, uh, how it relates to blends?

17:22 And this is, um, a long time ago. It's kind of, it's almost ironic now that Tableau's gotten so good at data modeling in the, in the age of AI, and why is that so important, um, for agents, et cetera, to be able to give all that kind of semantic context that you need. Um, but Tableau didn't start that way. They actually started on the exact opposite extreme.

17:40 There was, um, uh, the big BI vendors of the day were really Cognos, where I worked for 10 years, and business objects, um, that kind of forced you to have complex data models, which kind of got in the way. So Tableau really started without any data model at all. Like you just, um, you connected to data inside Tableau Desktop, and, um, and it forced you to join all that data together if you wanted to bring together, uh, multiple tables or multiple CSVs or Excel sheets, um, into one big table, kind of like a big denormalized table for anyone who understands data structures or like one view in a database, right?

18:15 And, um, a problem… We'll, we'll look at the one that came along with this. So sometimes you'd get one of these big data sources, and then someone would come along and say something maybe simple like, uh, you know, "I'd really like to be able to combine…" Um, you know, finance comes and says, "I want to be able to combine some of these targets we set out," uh, into…

18:34 And it'd be tricky because you'd have to bring it to the same level of detail. So, um, that would explode your data, and it could end up giving you wrong answers, and that's why people had to get good at LODs like, uh, uh, Jessica was talking about. I used to- Like, be I've been using Tableau pretty consistently since 2017. I used to love the fact that I was great at LODs, but now I almost never have to write them anymore, um, because of the data model in Tableau.

18:59 But anyway, so data blending came and let us join, um, two different data sources together. And then Tableau got a bit more serious, um, with Tableau Server and later Tableau Cloud, about reusability. So before this, every data source was embedded inside a workbook, which meant, uh, everyone who wanted to create their own set of dashboards had to start over and build the data source over again.

19:23 They couldn't share data sources between workbooks. Um, uh, but this came along in, uh, 2012. And I think at that time, people expected a lot of extensibility and being able to do things like keep data its natural grain and stuff, but that didn't come until, um, what c- has a lot of different names. Sometimes we refer to as the object model. But anyway, it's, uh, it's Tableau relationships, which came out a little over six years ago, and they, they allowed people to say, "I'm just gonna connect my tables together.

19:52 I'm gonna let Tableau worry about the joins at runtime, and that way I, I don't have to do things like complex LODs and count distincts and all these kind of slow calculations anymore, because my data sits at its natural grain." Um, and based… It's pretty crazy. This has been six years, and based on some kind of doing enough of these TUGs and informal testing, I think that maybe maximum 20% of Tableau data sources are built this way.

20:18 Um, they're still ba- built on that old kind of joining everything together, getting it flat in one table, which causes, um, massive performance issues. So I'd like to see more people do this, obviously. Um, and then, uh, Tableau added, it, it took four years, but what they added was the ability to add, um, multiple fact tables. So basically to allow you to share tables across, uh, different ones.

20:40 So what this lets you do is, this is where analysis gets really interesting, and it wasn't, uh, and it was very hard to do in Tableau before this, to answer questions like, "Hey, if our sales are going down, are they going down as a result of support calls?" Because to try to get sales and support calls, or maybe inventory or returns, all into, like, a single table is virtually impossible for anyone, you know, who's ever tried to do that before.

21:03 So Tableau now makes that pretty easy. And then, um, what was introduced now, um, is composable data sources. So I got that out of order. So basically, it took over a decade of these technical workarounds to be able to let you, though, take somebody else's data source, this is what composable data sources are really about, um, and extend that data source and make it yours, without having to open up theirs and risk all the many, many workbooks that might be connected to it downstream.

21:34 So, that's where composable data sources come in. And then, um, I wanna demo this and kinda why it's a, a big i- uh, a big thing. So first off, um, if we could just do a show of hands, which I guess I can't see, but, or put in the comments, I could look at it at the end or, uh, or when Kevin's presenting, is, is how many people do blends today, and do they find them a pain, I guess, would be…

21:58 That'd be great if people could, could throw that in. And again, we'll check it at the end, just 'cause it's hard to see that as you're presenting. But, um, here we are in Tableau Desktop. I just wanna show you. Here's, um, I built this data model, um, again, using relationships, on the Tableau bookstore database. So you can get this on Tableau.com.

22:16 You could just do a search, bookstore, um, Tableau. But basically what it is, this is a published data source, so I already created. It's on, um, Tableau Cloud, which you can see here. Um, and basically what it is, is it's got four quarters of sales. These are four union tables together. Um, and then books are sold by edition, right? And then editio- books have editions, and it's editions that get sold, and editions have publishers.

22:40 And then books have all this other dimensional information, like authors, awards, um, info, and ratings. Um, and what's neat about this model, if we look at it, is you'll see I already created this super, super simple viz, which is what do our sales look like, uh, over time, and break it down by genre. So, um, what's neat about that data model is that data model's not automatically exploded.

23:03 Tableau will dynamically, um, just use the tables it needs, um, uh, as it needs them to build the viz. So you can always come in Desktop or in the web, either one, if you're using web authoring. You can come to Worksheet, View Data Model, and what we'll see is if you come up here and go, um, Hide Unused Tables, Tableau will show you which tables it needed to do to draw this viz, right?

23:26 So you'll see that, um, uh, that this is broken down by genre, and genre comes from the info table. This, this dialogue will get better over time, uh, I, I've been promised um, by the product management team, right? But just to show you what I mean, if I, if I come over here and highlight genre, you'll see info highlights. So we'll see where it came from.

23:46 And if I highlight sales up here, we'll see where it comes from. Also, kind of a neat trick, if I pull genre off, um, those tables will all disappear because Tableau does, only needs to query that one table to be able to do that. If I put genre back in, it'll build this out. So that's just kinda neat, just so you know it's there if you ever wanna understand, um, how this is working.

24:06 Um, but today I wanted to show blends. So imagine we have the sales data, uh, we present it to an executive team, and they go, "This is great, but we really need the context of, um, uh, of how to, um, you know, to add sales targets to us." And of course, they send us sales targets like this in Excel, right? And it's this awful kind of, uh, crosstab-y looking thing.

24:31 You see them in Excel, like dates up here. There's no measure on it. Do you know what I mean? But there's how many books they expect to sell, uh, based on genre by, uh, by month, right? Again, it's only one year's of data to keep it simple. Um, so what I'm gonna do is I'm gonna duplicate this as a crosstab, a cool feature a lot of people don't use.

24:50 So here's our sales, uh, by month, and now, uh, how do we bring in, um, that, uh, that data? So in the past, it was really kinda tricky to do this because- Oh, took that too far out. Uh, we'll say like that. That's good. Um, so in the past, this would be really tricky. We would have to use a blend. The reason why is because before composable data sources, I can't go add that, um, in to the table.

25:19 I can't add my, uh, um, Excel into the table 'cause, hey, I don't own this data source, and I certainly don't wanna go back to server, bring this down, add it in, and then everyone else who's already using this data source, um, you know, I, I, I don't want them to have to be messed up by now they got these targets in it, right? So what you would have to do previously, um, to get a blend to work is you'd come up here and you'd have to add another data source, right?

25:45 And I could go Microsoft Excel, and then I could do that book targets, right? And it's gonna bring me back here to a data source. Um, here's a cool Tableau feature a lot of people don't know exists. Um, Tableau hates data in this format. It's very tricky, but I can click on those 12, uh, columns and pivot them, and then I could rename them, and we could call this Month, and we could call this Sales Target.

26:09 So Tableau likes, um, data longer than wider, if that makes sense. So it'll like every column, uh, to uniquely be its own dimension, if you wanted to think about it that way.

26:20 Um, and so now we have this data, and the question becomes, uh, how do we combine this data, um, with this, right? So what you'd have to do is say, um…

26:32 We could come here, and you'll notice that, um, I can link, um, Genre. I think it does it automatically. So by default, it automatically links it, right? Um, so what we could do is we could take Sales Target, and I could drop it in here, and now we got our sales target. But if you notice, these all look wrong. It's the same number all the way across.

26:52 Because of course, um, what we have to do is we need to be, uh, we need to have that blend work on a couple of different things, right? 'Cause I also have to have it work on month. This is just giving me, um, the target by genre total, right? It's giving me the sum of the target across all, but I want it at the month level. So there's a number of ways to do this.

27:15 Uh, we could come up here, go up here, and we could edit blend relationships, and I, I… This is kind of tricky to me, to me, so I'll even spare you that, which is a little tricky. What I could come back here though is, um, if you remember from this data source, um, uh, whoops, the other data source. Um, if we came back to the other data source, our targets one, um, we have Month as a name, right?

27:38 So what I could do is if I come back to the sheet, the sheet we were working on, uh, and I come back to our book sales, what I could do is I could create a calculated field called Month. I have to give it the exact same name and, or, or it gets a little tricky, right? But I could say Date Name, uh, of Month. Play around with this. Of Sales Date, right?

28:01 Um, which is, why did that, Sale Date, sorry. So this will return the name of the month literally, right? And it's gonna call it Month, and now I jump over, and I can also link on Month. And if I link on Month, um, then, um, it'll work, right? So I, I find it a little bit in- unintuitive how to do that. And another thing about these blends, if you notice, is they're, um, they're unique, um, to every sheet in the workbook.

28:31 So these are not now connected. So if I were to create another sheet, right, I come back here, they're not connected, you'll notice. It's only on the sheet level. So if I wanted those targets again, and you know we're gonna put together a dashboard, and I want it to show this is a bar chart. Uh, or sorry, s- b- side-by-side bars or bar and bar, or all kinds of cool ways you could show this, right?

28:52 Um, I would have to, on every sheet I compared them, I'd have to go back and hook all this stuff up, right? So, um, it's, it's pretty flaky, which I guess I showed it here and made it seem not that bad. But like, it's very flaky, and then a, a big thing I'll say about it is, uh, I was recently talking to one of our clients, like Dennis Mentrum, who's, uh, like a boutique Tableau consultancy.

29:15 Um, and they had this workbook that was taking, um, we timed it. It took three and a half minutes to open on Tableau Server, but, you know, could've been Cloud. And I, first thing I ask obviously is, "Do you have a blend relationship in there?" And they go, "Yeah, we have one blend, but it's not a big deal." And I'm like, "Okay, we'll see." And we went in and just clicked the little blend thing off, and the workbook opened in two seconds.

29:37 Like, I wish that was an exaggeration. It was three and a half minutes, and it was two seconds. Because, um, and we'll get into this a little bit, but we don't have a lot of time, of the way Tableau, um, generates this, um, is it does both queries, and the way it stitches together in memory at the end, they have to be at the right. Um, it's very easy to break this.

29:56 Like, if I were to, say, drill that down to day, like it'll break again 'cause remember, it's only at the level of month. Those kind of things. There's all kind, all kinds of things. So anyway, now it's much easier to do this, and this also, um, allows me to, um, show off composable data sources, which were here, and also why multi-fact, uh, relationships or, you know, the ability to share dimensions is so important.

30:20 So, um, what, the way composable data sources work, uh, and this allows us to do, um, a lot more, right, is what we can do is, this is something we could never do before, but now I can add a composable data source. So I've already added that Excel spreadsheet as a data source, um, inside Tableau Cloud. So now what I can do is I can say, "Let me go find this book targets," right?

30:46 Um, so book targets is here, right? I can connect to it. Uh, switching many columns of contents worksheet. Uh, let me not do that just for a second. Let me go here.

30:58 Uh, and then we're on book sales. Let me do it here. Just, I don't wanna… See, that might have broken my blend is how, um- Uh, is how flaky they are. So if I go to book targets, I shouldn't get that error message anymore. So I come to book targets and connect.

31:12 Uh, I did, but anyway, it'll be fine.

31:15 Um, okay. Now what I could do is… Remember, so this book has genre in it, right? So one thing I wanna do is, um, I wanna be able to take this, and genre is in this info table. So what I do is, now it's just Tableau relationships for anyone who know it. Like, it's a lot better. I'm extending this data model by composing my own data model, which is where composability comes from.

31:38 So I say, the relationship I want between these, um, is based on genre. So all I have to do is go genre-genre to link them like that. Now, it'll be available to me. Now book targets are available to me, um, anywhere I go, right? But remember, I want a better way to link dates together here as well, right? Because this has dates and sales has dates.

32:00 So right now both of these tables are sharing genre information, right? But the next thing I wanna be able to do is I also wanna be able to share dates. So what I would wanna do is I would also wanna have a date dimension here, um, that I could hang everything off of, right? So, uh, I'll come in here, and I already have a date dimension, um, that I already uploaded.

32:22 You'll see it's certified here. Um, so it's certified by me. It's an d- official date dimension for all da- data sources we have. I just add this in, right? And, um, again, I could've taken the blend off first. Um, but now what I wanna do is, they don't have to be in this position. They can be connected this way as well, right? So sales to date, right.

32:45 Uh, and remember, the, the neat thing about this is they all get to stay at their own level of detail, right? So I can come here and come down here and go sales date maps to date, right? 'Cause we have sales pote- many sales potentially on every single day, right? Is the grain of this, is at the level of day. Um, but this one, if you remember, was at the level of, um, this one's at the level of month, and then I'm almost done here.

33:11 So then what I'm gonna do is I'm gonna combine combine with, again, same kind of idea here. There's an… I could have converted it to a date first, but I'm gonna turn this into, um, like that. So they'll be able to join at a completely different level. So this is join… Or sorry, this has a relationship, and we'll join at the level of day, and this at the level of month.

33:34 In any case, what happens now, which is kind of neat, is now my book sales, right, I could pull date up here. Right, well, let's do it this way. I could control, I could control it up, bring it up this way and say I want date like this, right? And then if we watch again, I could keep this open. View data model. Um, again, if we hide unused table. So right now I, I'm only bringing in, um, uh, from that date dimension, right?

34:00 And then what I could do is I could bring in, um, our sales target, and we could put this in, and you'll see that this will dynamically show that now we're using two, right? And then what I'm gonna do is we wanted to break it down by genre, right? And that will show the other table I need to get. And then the other thing we wanna do is we want a count of our sales to compare it, and you'll see how Tableau is dynamically smart enough to go get, um, and make the joins it needs to be able to do to bring that data in.

34:31 Um, I'm pretty much out of time, so just the last thing I wanted to show you, which is kind of neat even, is if we created a calculated field to show… And, you know, we could create a heat map or something if we had more time. But if we went sales ratio to target, terrible name, but then I could go, we could take sales and divide it by, um, the target to know, you know, if we were on as a percentage.

34:55 What's kind of neat about this is if we now take this and, say, also put it on text, right? We could make this a heat map, right? Um, so this is going to be a percentage obviously. So a percentage…

35:11 Oh, I must have done that calculation wrong. But anyway, the neat thing, someone could catch that in the replay. But the neat thing is, um, had I done that right, when I highlight on… Whoops. When I highlight on the sales ratio, you'll notice that, um, I'm getting, it'll highlight that it comes from multiple tables, which I also think is kind of cool.

35:30 So again, I was kinda, whoops, rushing on the, uh… Oh, I know what I did. I just, uh, I changed the wrong ones' formatting. Sorry, 'cause I put it there. Just a sec. Format and then percentage. Uh, and they'll receive the percentage. And again, we could highlight this based on the percentage to get a highlight table or however we wanted to see it.

35:51 But again, just kind of neat on how it shows where it comes from.

35:54 Um, so I'm gonna just wrap by quickly saying that what this basically comes down to is why blending was always a compromise. Um, performance was a super big thing about this. Um, the, there was always this importance on you had to know which was the primary and which was the, uh, secondary source of data, which became very confusing for it, right?

36:16 The aggregation happened before combining, um, which made things, um, messed up. It was really the fragile linking though, 'cause you could break the workbook very quickly by pulling the wrong kind of field on, right? And then there's a lot of calculation filters don't work like they would work with a composable data, um, with a composable data model.

36:36 And, um, of course the integration logic lives inside the viz, right? And specifically, it resides inside every sheet it's on instead of in the data source, so no one else could reuse that. They would have to build that every single time on every single sheet. And the other one, if we publish that data source, is a new published data source composed of the multiple ones, is that anybody could reuse it.

36:59 So, um, that's my wrap. I'm probably two or a couple of minutes over, so I don't wanna Uh, dig into Dennis' talk, uh, Kevin's talk anymore than that, so. But I think w- hopefully we have time for Q&A at the end. If not, uh, you guys know where to find me, for sure.

37:14 Dennis Kao: Thank you, Kirk.

37:17 I think we're gonna y- have a new term. It's called Kirking, which is da- you know, synonymous to data modeling, right?

37:27 All right.

37:28 Um, if you have any questions or, um, please use the Q&A option.

37:34 Um, I'm hoping that maybe toward the end we'll have some time to address questions live. But, um, in the meantime, Kirk will be checking the chat as well as the Q&A.

37:44 All right. So next up is Kevin.

37:48 Um, Kirk is a tough act to follow, but I think Kevin, you know, can hang.

37:52 Uh, you can read his bio as well. He's currently the lead, a lead bio- business intelligence analyst at Capital One.

38:00 Um, and you can see all the kind of accomplishments. Um, most notably, he was just, just on Iron Viz as a finalist this past year.

38:11 Um, he has 10 Viz of the Day. And, um, and I think that is particularly notable and highlights kind of his design and, um, one of his design skills. And y- w- if you go to his, uh, Tableau Public website, you'll, you'll notice that, you know, he is very, very skilled at design. Um, I think, uh, he has, uh, go, definitely go check out some of his videos and presentations that he's done.

38:46 He uses Figma, um, and has demonstrated those in videos. And, um, we're really happy that he can join us and showcase some of his, uh, give us some best practices and tips.

39:00 So I'll turn that to you, turn to you, Kevin.

39:03 Kevin Wee: Sure. Thank you for the introduction.

39:06 Let's get started here.

39:09 And I'll be loo- I'll be checking on the group chat as well, so if anyone have any question that I can address right away, I will do so. Well, uh, given that you have made such a great introductions, I will speak less about myself. But just to know that today I'll be talking about the applications of color in, um, Tableau dashboards. And for the context of today, um, presentation, I'll focus on business dashboards other than the ones that are more flashy, more infographics.

39:39 So with that said, um, my Chinese name is Huang Kewen, so if we have any Chinese audience today, just to make a connection.

39:47 Um, yes, it's all about me. Done, done, done. Okay.

39:51 Now, let's talk about, um, my pre- my presentation today. I'm gonna start by talking about several color attributes. We're gonna talk about how to apply color in dashboard formatting and chart formatting, and then we're gonna talk about how to use it to encode data in our charts.

40:09 Um, if we have more times, I'll talk about Tableau's new functions regarding colors, and I'm gonna share more open non-Tableau resources with you so that you can better tweak the color in your business and analytics dashboards.

40:24 By the end of today, you should be able to at least describe three color attributes. You should be able to at least identify three non-data visualization component in a dashboard whose color you may format.

40:39 You should be able to discern the applications of four different families of data visualization palette.

40:46 You should be able to identify at least one newer Tableau color function.

40:52 Um, with that said, let's get started.

40:56 Um, people often describe color in a very vague way, right? Make the color pop. What does it mean? Or, um, make something lighter, darker. Do we have, uh, an agree, um, have a definition that we can agree on? So here are a couple of, um, commonly used, um, color attributes that may help you when you are designing a dashboard with or without a business partner.

41:20 Number one, it's transparency. That's basically a descriptions on whether you, how much you can see the element underneath that color mark, right? Something that's more opaque would be something that's not transparent. Something that's transparent would be something that's not opaque at all.

41:38 Um, hue, it's basically if you see the typical color wheel, hue is basically the angle in which, uh, a color lies in the wheel. And then the typical color hue would be, you know, uh, whether something's red, it's pink, it's purple, it's teal.

41:54 That's different hue.

41:57 Saturation is telling us, uh, conventionally speaking, it's how much pigment is in the color.

42:03 Digitally speaking, it's how far away it is from a gray equivalent in the same, you know, saturation, um, in a, in the same lightness. You can see that in something that's more saturated, it's more blue on a screen, and something that's less saturated, it's basically gray.

42:21 The other attribute, it's how light or dark a color is. That's basically telling us how close a color is to white or black.

42:31 Um, so this is lightness.

42:33 Um, now I'd like to talk about color contrast, which is very important. Color contrast, it's whether you can discern two color elements on the screen.

42:44 And you can see that here I leverage, um, the difference of lightness to make a good contrast in color. And you can see that, um, I use lightness instead of saturations or hue. And you may ask why contrast is about lightness, not about other, um, color traits. Um, let's look at this chart here.

43:11 I would like you to, within five second, count how many strawberries is on the screen.

43:18 They all have different color, different hue, but the same lightness, almost the sa- uh, the almost the same lightness, uh, the same saturation.

43:33 Anyone wanna put in the chat how many strawberries that we have?

43:41 Can you feel how much your brain is hurting when you're trying to do the math?

43:46 How about now?

43:47 Without any hue, just using lightness and saturation, how many strawberries that we have?

43:56 Is it easier to find strawberries even though there's no red color?

44:03 Of course, in real life, we leverage both color, saturations, and lightness to find objects, but I guess the message I'm trying to get across it's that if you, um, in real life, most of the times when we discern different objects or trying to count things, lightness play more roles than hues.

44:27 So don't, uh, spend that much time tweaking the hue. Most of the times it's all about tweaking the contrast and lightness in your dashboard.

44:36 So how can we apply everything that I've just said in a business dashboard design?

44:43 Let's look at, um, this worksheet. Note that it's currently on my Tableau Public, so later you can go, you can download the works- workbook, take a look at, uh, what's inside, just so you can learn from my content.

45:00 I'd like you to look at this scatter plot, right? Typical scatter plot comparing the profit and sales of different orders.

45:09 Um, how can we leverage what we have just learned? Number one, we can increase, um, the transparency of the mark so that when things are cluttered together, we can see how many… we can do some calculations on how many, um, things behind it.

45:25 So by reduce the mark a little bit, you can see that now it's better, easier for us to see. And this is not about color. You can actually also change the shape to open circle.

45:34 Easier for you to pinpoint specific mark.

45:38 Um, you can also tweak the contrast of the line and the text on the charts. For example, like this.

45:48 Now you can see that by increasing the darkness of the text and creating some contrast between the grid lines, axis, and the zero lines, the chart's easier to be read. So this is an improvement that you can make in your day-to-day business dashboard.

46:06 Um, a lot of times I've seen bar charts like this. People insist to, "Oh, we have a metric. We're gonna slap color on it." But the color doesn't really add much to the chart. I understand people like to emphasize how much or little a bar is, but a bar chart, it's doing that just fine. If you really want to do color, I would recommend you to basically just color, um, the region between the line and the zero line by using a s- very simple dual-axis area and line chart.

46:36 And then the area, you can just reduce the opacity, like this. So something more fashionable. You can easily find the maxima- maximum and minimum value in the chart.

46:48 Um, you don't have to always use a gradient of color in your dashboard.

46:53 Now let's look at this use case.

46:55 Now we have, um, um, segments in detail card. So each line represent each segment. In this case, of course, you may use categorical color palette in your design.

47:11 And as a matter of fact, categorical color palette is not the only type of data visualization palette you may use. Let me go back to my presentation.

47:26 Oops.

47:29 There are four types of, um, color palettes that you may use in your business dashboards. I've just mentioned categorical color palette. These are the color pal- ca- uh, these are the palette that basically each color is its own hue. They may discern a little bit in saturat- lay- saturations and lightness, but mostly you're discerning marks, uh, by hue.

47:56 So this is one example. We typically use them on discrete values.

48:02 Sometimes we also only use two kind of color, basically gray out the rest, only h- emphasizing or alerting a specific category.

48:11 Um, usually if something that's good, something that's worth to be celebrated, it's blue or green. Something that's concerning, it could be yellow, red, or orange.

48:23 For values that are, um, continuous, such as, um, sales, such as quantities, we use sequential color palette. Basically, this is the palettes where the color go from very highly saturated light color to very, uh, less saturated dark color.

48:43 It could be within the same family, such as different shades of blue.

48:47 You can also have a multi-hue sequential color palette that go from one color to another. Usually a good design of a sequential multi-hue color palette, it doesn't go in a wide range in a color wheel. I'll explain why later. But usually they just go a little bit, like from yellow to green, or yellow to green, and a little bit to blue.

49:09 Um, if you have- Continuous values that cross zero, such as profit, such as, um, yeah, such as profit or whether a shipment is late or devi- or delivered early. Um, a diverging color palette, which is a combinations of two sequential color palette, uh, where the gray part or the white part meets. Um, you can see that usually this, the color on both ends, they have high contrast in hue.

49:38 For example, orange and, um, blue or red and black. They don't have to be complementary color, but they just need to be, uh, different enough for people to simply tell, um, them apart.

49:50 So going back to my present- my workbook again. One, two, I believe it's this one. Nope, that one. Great.

49:58 Now, um, another, um, typical applications of, uh, categorical color are for to discern Boolean value. For example, on a screen here, we have the sales, um, from other Canadian cities, uh, versus Vancouver. And people are like, "Okay, we're gonna use the default, uh, categorical palette on Tableau." But the thing it's usually people only focus on the true value, not the false value.

50:24 So maybe you can tweak this by, instead of emphasizing both color, you can just emphasize the orange and change the blue to gray and then move it to the back. That is a way to use a categorical palette in a boo- uh, for Boolean breakout

50:41 or things, cases like this. Uh, people love using traffic, uh, cr- uh, traffic light color palette in things that they are trying to make a judgment on something that's whether good or bad. Um, but here's the thing about a traffic light color palette. Let me show you what I mean here. Let me bring it here. So this is a traffic light cone, right?

51:04 And if I apply a simulator for a specific color blindness, that's red-green color blind, you can see that it's really hard to tell apart which one is green, which one is red. For people with cl- color weakness, um, they're just, just different shades of yellow. They can tell the difference between blue apart though.

51:25 So how can we leverage this, um, nature and make our design more inclusive for different people?

51:36 You can shift the green to cyan. You can see that I just use the same color palette for the rest.

51:43 Um, shift the blue to gree– I'm sorry, shift the green to cyan, and then darken the red a little bit.

51:50 Now you create a contrast not just in hue, but also in lightness.

51:56 So this would be a better way to use the color to show judgment in your breakdown.

52:03 The other thing, it's sometimes, um, the breakdown is actually… Sometimes the breakdown is just different, um, point of a continuous value, right? This is even though they are look, they look discrete, but they actually have a clear order. So in this case, using hue doesn't make sense because we know that in a rainbow, blue doesn't go to orange, and an orange doesn't immediately follow by red, and then go back to, uh, teal.

52:31 So here's what you can do. You can use sequential palette, but apply it in a discrete setting. So in this case, there's a benefit for this kind of design. That is, you can quickly tell the outlier apart. For example, in, um, September and March, we can see that 2022 is clearly, um, second highest, even though the rest of them follow a trend. So using a sequential palette make you easier to tell things that are out of order.

53:03 Or, uh, in this case, you have a, um, ordinal category, like different profit ratio, and then one of them actually crossed zero.

53:13 In that case, you can use a kinda diverging color palette. Um, the positive one is blue, and then the negative ones are different shades of, uh, orange to red. That would be a way to do, uh, to apply color as well. I have a couple of use case. I'm just gonna go through them quickly. Um, this is another use case. Um, you have different classes, but one of them is outlier.

53:38 For example, you have first class, second class, standard class, but then you have same day. Clearly, this is not, um, in this order, right? It's not, uh, with order with the rest. So what you can do here is you can apply the rest in a sequential palette and then just gray out the outlier. Just give it a gray color if it's not a big deal.

54:00 Lastly, we all love rainbow color palette. Uh, Tableau has dysfunctions which I argue sometimes do more harm than, um, good. Um, people love applying it, uh, when they have a large, uh, number of, what do you call this? Um, breakdown values. The thing about a rainbow color palette is this.

54:21 Oh, not you.

54:25 Rainbow color palette from the perspectives of someone that's completely color blind, um, they have different perception illuminance, meaning that some color just inherently brighter than other color. For example, we see yellow very bright, but blue and indigo just inherently very dark, even though you tone out, uh, the, the, the brightness, uh, max out of brightness.

54:49 The same go with the versions of the color palette in Tableau.

54:53 That's why it's not a good, um, uh, applications in a dashboard because

55:02 you would think that the brighter has more data value than the darker one. And also it's very hard to tell people, what do you mean by the yellow line or the teal line? Which is the teal line, right? It's very risky for you to use k- this kind of color palette. You will risk a data misinterpretation. Um, the solution that I offer to my partners often just to Create some kind of selection mechanism so that you can highlight a specific line and then gray out the rest, because you won't be able to read it better than this way anyway.

55:34 So now let's look at, at the dashboard value, uh, dashboard level. So in terms of business dashboard, you can see that when you put a chart together, even though the, even though they arrange very beautifully, it would be better if you give the container a very thin line border and then the background just a little bit gray color, just to create more contrast to group the containers.

55:59 Um, these are all just to apply a color on a very thin layer of the container or at a background of a container. So these are different ways for you to apply color as well.

56:10 So,

56:12 okay, I have one more minute, so I'm not gonna go too much into this, but, um, the two new Tableau functions. Um, so starting last year, Tableau has two ways for you to play with colors. Number one, you can directly add color, um, in the create custom palette function. You can click on it, and then here you can click the pa- paste swatch. So let's say if you have some color palette here, you can just grab them.

56:43 Um, they're all, you know, with the ha- pound sign out front and then separated with commas. You can just copy, paste, add, and then now you can add a new color palette. You can call them whatever you want. Um, one, two, three.

56:59 So this is how you add, um, categorical color palette. Uh, the same way you can also, I think now it's here. One, two, three. Yep.

57:09 Um, you can also add color to, I believe, something sequential. Let's go to

57:15 here. Okay, cool. Let's add the color.

57:19 For something sequential or, um, divergent, you can also add a color palette here. Create custom color palette.

57:27 Grab the color that you have. Where is it? One, two, up.

57:31 You, you color. Oh, I couldn't find it.

57:37 Okay, here. Okay.

57:38 So here are the color.

57:42 Copy, paste.

57:45 You can get rid of this. You can select whether the palette is sequential or diverging, and then you can save it as well. Yes. So these are different ways to add color. The other thing that you can add color is, let's say in one workbook you have very diligently, you know, format all of the, um, charts.

58:06 You have set this all up very clearly. Here you can export the custom theme as in the JSON file. And then in a new dashboard, let's say in this dashboard, I can actually apply the theme.

58:22 Um, import theme.

58:26 Okay. JSON file.

58:30 Overwrite.

58:31 And immediately you can change the color and formatting of every charts all at once. So these are two new very powerful Tableau functions.

58:40 Great.

58:42 Okay. So I'm not gonna spend more time talking about open resources. I just need you to know a couple of things. Um, I will share them, uh, in the link in the chat later. So let's wrap things up together.

58:53 We have covered color transparency, hue, saturation, lightness, contrast, and a little bit about perceptual luminance.

59:01 We talk about how you may format the mark, the text, the line, and containers in every part of the business dashboards.

59:09 We talk about, uh, the applications of four different categories of color palettes: categorical, highlights, sequentials, and diverging.

59:18 Um, we also cover, uh, two new Tableau functions that are custom color para- palette generators and import and export, uh, color themes. Um, that part you can find it on my website, kevinwee.com.

59:32 So that's it. Sorry for going, uh, off time for three minutes, and open to questions.

59:40 Iris Sun: No problem. Thank you so very much.

59:43 We do have some question, right, in the Q&A.

59:47 And thanks, Kirk, for ha- uh, have, uh, Kirk has already, uh, just some of the questions, and I think we have some new questions for Kevin. Is that right?

59:56 Uh, one question from Jessica.

59:59 Jessica, do you wanna ask your question?

1:00:02 Jessica Bosch: Yes. One sec. Where is my question? Ah, yeah. Um, so I wanted to ask you about when you design colors, um, for instance, and specifically that are also accessible.

1:00:13 Kevin Wee: Mm-hmm.

1:00:14 Jessica Bosch: I wonder if you have a certain process or tool that you use.

1:00:18 Kevin Wee: Oh, yeah. Sure.

1:00:20 Uh, those are actually, uh, the closing, uh, part of my talk, uh, about sharing some of the tools. So, um, earlier you saw that I was using a plugin, right? This eyeball basically is a simulator for different, um, versions of color blindness. The most common color blindness, uh, it's, uh, green-red blind. Basically, these are the people that cannot discern, um, green and red.

1:00:44 And then, uh, the most severe one, of course, is the one that you can only see one color. So monochromacy. Yeah. So I like about this, um, simulator is when you have different color palette, you can just click through them, and then just to see whether other people can see the color well enough or not. So this has been great. Um, I also use the Adobe Color Checker and Adobe, um, palette generator a lot.

1:01:07 Like this is for you to, if you want to have a good, uh, complementary color or different combos of color, you can use it to find, um, a good starting point for the palette.

1:01:21 I hope this answer your question.

1:01:23 There are one more in my blog.

1:01:25 Iris Sun: Yeah.

1:01:25 Kevin Wee: Everybody go check it. Yes.

1:01:26 Iris Sun: Oh.

1:01:27 That was helpful. Thanks.

1:01:31 Kevin Wee: Yes.

1:01:31 Iris Sun: Yeah, we're a little bit behind the schedule. Are we okay to stay for another five minutes for question, for Q&A?

1:01:39 Okay. Yeah. So I think we have one more question from Srini. I hope I say the name correctly. Um, so the question is for, uh, Kevin. Uh, so again, uh, excellent, uh, sections. The question is: How important does color remain since November 2025? What happened in 2025? November 2025.

1:02:02 Kevin Wee: November… Wait, what happened November 2025?

1:02:06 Iris Sun: Yeah. So that, that was the question.

1:02:10 Kevin Wee: I'm sorry, I don't… Uh, I guess I don't understand. Okay. How important does the color remain, uh, since…

1:02:17 I guess I don't understand the question. So if Srini, you can come out and elaborate, I…

1:02:23 Maybe I can better explain it, or does anyone else understand that question? Can someone explain that? How important does the color remain since…

1:02:33 Iris Sun: I can't find it. Maybe he left already. I can't find-

1:02:35 Kevin Wee: Yeah

1:02:35 Iris Sun: … I tried to make…

1:02:36 Kevin Wee: I'll, I'll see

1:02:37 Iris Sun: … I tried to make him, because I was not able to find him.

1:02:39 Kevin Wee: Mm-hmm. Yeah. I'll reach out-

1:02:40 Dennis Kao: I, I think they were referring to one of your chart, your line charts, I think. Um…

1:02:44 Kevin Wee: It's like what… Okay. The, the one with the…

1:02:48 This one?

1:02:49 Dennis Kao: Not sure.

1:02:52 Kevin Wee: This one? Oh, sorry.

1:02:54 This one?

1:02:55 Dennis Kao: Yeah, I'm not, uh, yeah, I'm not sure. So, like if-

1:02:57 Kevin Wee: Might have been-

1:02:58 Dennis Kao: I don't know if, if Srini's still on, they can expand in the chat

1:03:02 Kevin Wee: …

1:03:03 another one.

1:03:04 Dennis Kao: It, it might've been an LLM related joke, not joke. I'm not sure. Just ba- based on he weighed in on. I said I want a, I want a Kevin Wee color, uh, markup file, but I'm serious about that.

1:03:17 Kevin Wee: It's okay. We can… Uh-

1:03:18 Dennis Kao: Markup file

1:03:18 Kevin Wee: … I'll reach out to him, uh, on the side and see, and-

1:03:21 Dennis Kao: Yeah

1:03:21 Kevin Wee: … help him out later, but-

1:03:22 Dennis Kao: Yeah.

1:03:23 Kevin Wee: Yeah.

1:03:23 Dennis Kao: And, and Kirk, is, is there any responses that you wanna maybe expand on that might've been difficult to answer, uh, in writing?

1:03:32 Kirk Munroe: Uh…

1:03:33 Dennis Kao: Since we're, since we're hanging out.

1:03:34 Kirk Munroe: I don't, I don't… I think they're pretty clear.

1:03:36 Dennis Kao: Okay.

1:03:36 Kirk Munroe: Other than Jessica's, unfortunately, there's no easy way to replace blends now with composable data sources-

1:03:41 Dennis Kao: Mm-hmm

1:03:41 Kirk Munroe: … 'cause the way blends were created, but that's a bit of a drag.

1:03:45 Iris Sun: Oh, wonderful. Thank you all for your, yeah, for wonderful presentation, and, uh, thank you all everyone for coming. And, uh, I, because we are already behind, so we don't want to spend too much time, or to hold you here forever, but we do have another upcoming event already scheduled on November 20th. I will send you, uh, uh, e- details in emails, uh, in a few days.

1:04:06 That's, again, thank you all very much for a wonderful presentation, and for all, uh, giving your time for, uh, joining us today. And that's all. Thank you.

1:04:17 Yeah.

1:04:17 Kevin Wee: Thanks for having us.

1:04:18 Iris Sun: Again, if-

1:04:18 Dennis Kao: Thanks for having us.

1:04:19 Iris Sun: Yeah. If you have any questions-

1:04:20 Kevin Wee: Thank you, everybody

1:04:21 Iris Sun: … send us email.

1:04:22 Thank you, everyone.