Tableau Workbook Optimiser vs Tableau Exchange Accelerators!

0:00Hey, it's Tim here. In today's video, we're

0:01back after a short break and we're going to

0:03be testing

0:04Tableau accelerators using the Workbook

0:07Optimizer test that Tableau released in the

0:09latest version

0:10of Tableau. So we're going to see how the

0:12Tableau accelerators themselves actually

0:14fare in Tableau's

0:15own performance testing capabilities. Let's

0:17get started. Now in order to do this test,

0:20we're going

0:20to need to set a baseline. We need to sort

0:22of normalize our sensors to see what the

0:24Workbook

0:25Optimizer scores something that you know

0:27has been built by Tableau as a default. And

0:29also we need to

0:30come at this as if you're a new user.

0:31Essentially you're going to be starting at

0:33Google, you've

0:34maybe heard about Tableau accelerators, and

0:36we're going to find what Tableau is pushing

0:38to you as a

0:39user. So in Google, I'm just going to go

0:41ahead and type Tableau accelerators, if I

0:43can actually spell

0:45that, let's say Tableau accelerate, you can

0:47see I've actually searched it in the past.

0:49So when you

0:49hit that you get this top link now. I've

0:51disabled ads so you might see other things

0:53being sort of

0:54pushed up and down here. But for me at

0:56least, the Tableau accelerators hit that's

0:58at the very top

0:59is this page, I'll go ahead and click on it

1:01and you'll see this particular page. This

1:04is a

1:04marketing page that I think is pitching a

1:06bunch of accelerators in different sectors

1:08by department,

1:09by industries, and then by enterprise

1:12applications, essentially tools like Sales

1:15force,

1:15ServiceNow, and so on and so forth. But

1:18before I get into these accelerators, what

1:20I'm going to

1:21need to do is to sort of have a normalized

1:23baseline. So the best baseline I always

1:26think

1:26to use is the default workbooks that come

1:28with Tableau desktop. So if I just go to

1:30Tableau here,

1:32I'm talking specifically about these three

1:34workbooks. So to get a baseline, I'm going

1:37to take

1:38all three of these workbooks, I'm going to

1:40open them, run the optimizer immediately,

1:42see what score they get, then I'm going to

1:44average that score to basically get my

1:46baseline score. And

1:47in essence, we want to try and see if the

1:49accelerators pass this baseline score. And

1:52we'll basically just see what we find out.

1:54So the first one, I'll open up Superstore

1:56sales,

1:57we'll get that open, and we'll get the

1:59score. Now for this, I might cut in between

2:01sections,

2:02so you won't see every single sort of

2:03chronological thing that I do just to make

2:05it

2:06better to watch on YouTube. So stick with

2:08me or you can jump ahead in the timestamp

2:10and catch

2:10something else that you're more interested

2:12in. So I'll go ahead to server here, select

2:14run

2:15optimizer, we instantly get a score, we get

2:1710 out of 12. So the first one got 10 out

2:19of 12.

2:20The one that failed is unused filled

2:21multiple data sources have unused filled,

2:24and it's just

2:24a couple of fields. So this, this to me is

2:26like a full pass in many ways, this should

2:29be 11, because

2:30you're never ever going to sort of fully

2:32remove all unused fields unless you're

2:34really, really,

2:35really, really in production. And it's, you

2:37know, been completely locked down. The

2:39other thing that

2:40needs review here is calculation uses

2:42multiple data sources. Now this isn't a bad

2:45thing.

2:45Sometimes you have to use multiple data

2:47sources. Let's say you have targets, maybe

2:48let's say that

2:49blending is the only way you can do

2:51something, or maybe you're doing a cross

2:53data, cross database

2:54join or anything like that, you might need

2:56to use multiple data sources. So again,

2:58this doesn't

2:58sort of worry me. So in many ways, this is

3:01a good pass, 10 out of 12, perfect score.

3:04So we're going

3:05to do the rest a lot, lot quicker than that

3:07. So let's go ahead to regional, we'll open

3:09that it

3:09actually chucks me back into the existing

3:11workbook. And we'll go ahead and do that.

3:13It's actually

3:14opening in another window, it's gone to

3:16another screen. Let's get this on this

3:18screen here. Let's

3:19go to server run optimizer. Again, 10 out

3:22of 12. That's two 10 out of 12. This

3:26average might be

3:27really easy at this rate, because I'll just

3:28close that previous one there. We'll open

3:30up world

3:31indicators, this will open up in another

3:33tab as well. Let that run. And once that's

3:36done, we will

3:37get that going. Yeah, that's open on

3:39another screen. Here we are, this is

3:41actually got a

3:42couple of sheets on its own. But it does

3:44have some dashboards. If you look at

3:45tourism, business and

3:46global indicators, then that's all there.

3:49So we'll go to server run optimizer and

3:52drums 11 out of 12.

3:53So our average if I get my calculator out

3:56is going to be 10 plus 10 plus 11 divided

4:00by three, 10

4:01plus 10 plus 11 equals divided by three,

4:05the baseline score is 10.3 out of 12. So 10

4:10.3 divided

4:11by 12 gives me 86%. So the pass mark for a

4:15good workbook is 86%. I'll use the

4:18percentage as the

4:19past month because that's easier when we've

4:21got different scores sort of running

4:23through. And

4:24it'll also be easy to sort of get a general

4:25average for each of the sections that we

4:27saw

4:28previously as well. So the baseline scores

4:3086%. That's essentially 10.6 out of 12. So

4:34we're going

4:34to go ahead with that we're going to close

4:36all of these. And what I'm now going to do

4:38is access

4:38accelerators like I'm just a newbie. Let's

4:40say I've heard about these accelerators

4:42that Tableau

4:43Cloud can offer. And I'm just going to go

4:45ahead and go to the Tableau accelerator

4:47page here,

4:48I'm going to scroll down and I'm going to

4:50probably take two out of each section. Okay

4:53, so I mean,

4:54I could say here going through all of these

4:55, it's actually quite tedious, because there

4:57is no API

4:58at the moment. That's an idea on the ideas

4:59page, I'll put a link to that in the

5:01description,

5:02please vote for it. And but essentially,

5:03there's no API. So I can't just upload

5:05these all to Tableau

5:06server and then get a score coming back,

5:08that would be a really, really nice feature

5:10to have

5:10it be able to do that. But nonetheless, I

5:12'll just download some of these and then we

5:14'll run the

5:15score. Now, I'll make a general point about

5:17these later on, if you haven't already done

5:20so look at

5:21my previous video where I talk about

5:22accelerators, you can see that I've been

5:24rather quite negative

5:25about them. And I think this will probably

5:28be a good video to sort of try and normal

5:30ize that a

5:30little bit using Tableau's own tools to say

5:32, Okay, where could these be improved and

5:34start to try and

5:35be constructive rather than be 100%

5:37critical. So let's go ahead, let's start

5:39the very top

5:40departments is the first sector, let's say

5:43I work in sales, and I'm an expert,

5:45according to this,

5:46let's go ahead and click on that. What it

5:48does, it sends you to the Tableau exchange,

5:50Tableau

5:50exchanges like the ecosystem is the App

5:52Store for this particular setup. And you

5:55get a bunch of

5:56things, you get a description, you actually

5:58do get a version of the dashboard just down

6:00here,

6:00it's got quite a lot of tabs, a lot going

6:02on. But what we'll just go ahead and do is

6:04download it

6:05right from the top. And once we're in, we

6:07should get the workbook, and I'll create a

6:10new folder

6:10here. And I'll call this accelerators. Now

6:13I've actually previously downloaded all

6:15accelerators

6:16in the past. But we won't go into that now.

6:19We'll worry about that some other time. And

6:22just to make

6:22sure I'm using the latest and greatest of

6:25Tableau, I'll open up 2022. I've got such a

6:30weird thing to

6:30say 22.1. Let's just say it that way. And

6:33we'll bring it back, you can see the three

6:36previous

6:36dashboards that I've created. Let's just go

6:39ahead and close all of those. And what I

6:41will do for

6:41this particular case is I'll go to my

6:43desktop, I'll open up this which I can't

6:45spell, I'll drop it in.

6:47And this will mean that I am definitely

6:49opening up that accelerator directly. And

6:52when we get there,

6:53we don't actually need to go through each

6:55and every tab, there's quite a lot of tabs,

6:57a way you can see everything in one go is

6:59if you go to the very bottom right here,

7:01there's this

7:02thumbnail view. And actually, if you click

7:03on that, you get this view, which kind of

7:05gives you

7:05an overview of all the tabs in a nice place

7:08, a little hack, if you're in this view, the

7:11quickest

7:11way to get Tableau to refresh all the data

7:14sets is to actually right click on this,

7:16refresh all

7:16the thumbnails. And in order to refresh all

7:18the thumbnails, it has to ping all the data

7:20sources.

7:21So this is actually a really good way of

7:22refreshing all your data sources, making

7:24sure

7:24you have the latest and greatest data from

7:26your data sources. That's a free tip for

7:28you. But anyway,

7:29we don't care about that we're going to go

7:30to server. Actually, it's not gonna let me

7:32do that

7:33in this view. So let's go to one of the

7:35views. I've just selected view now it won't

7:38let me do it.

7:38Let's go to customer insight. Let's go in

7:40there. And now that we're there, we can run

7:42the optimizer.

7:43I'm curious why we can't do the optimizer

7:44in thumbnail view. That's a strange one.

7:46But let's run the optimizer and first score

7:49six out of 12. Okay, so let's see what this

7:53one has

7:53failed on. So unused fields. Okay. How many

7:56unused field? Okay, quite a few. Alright,

8:00so this this

8:01one has an absolutely massive data sources

8:04here with an absolute sort of sea of unused

8:06fields. Now,

8:07to be fair, if you're making a template for

8:10someone to download and get started with,

8:13you're probably going to leave all the

8:15fields available. Because if you hide them,

8:17then it's actually more work for the user

8:19to go in and find them and re enable them.

8:21So

8:22I will sort of give this a pass because

8:25actually that that is a good thing to do.

8:28And what you

8:29probably want to do is have some

8:30documentation to then tell the user, hey,

8:32according to our best

8:33practice, once you've finished designing

8:35this, remove the fields you don't need. So

8:38the unused

8:38field, it fell on that one now five need

8:40review. And so five need review. These aren

8:43't necessarily

8:44failures. These aren't bad things. These

8:45are just things to look at. But we want to

8:47look and see

8:47what they are numbers of views within the

8:49dashboard. So multiple dashboards contain

8:51more

8:52than 10 views. Okay, so all of these dash

8:54boards contain 10 sheets. Well, look at that

8:56top

8:57performers contains 21 shades. Now, it's

8:59strange because obviously, this is Tableau

9:01's own best

9:02practice. And the template has gone and

9:05sort of infringed on this multiple times, I

9:08would,

9:08I would say it's okay if you do this a

9:10couple of times. But when you've got it six

9:12times,

9:12and you're sort of breaking habit, you

9:14could start to maybe put together an

9:15argument that maybe

9:17these should be separate dashboards with

9:19interactivity between them. So you can go

9:21to

9:21those things. But again, in order to

9:23deliver this as one thing, it has to be in

9:25one dashboard. So

9:27this could be a workbook that's been

9:29designed around the deficiencies of Tableau

9:32itself,

9:33because for it to all be one experience has

9:34to be in one workbook. So that's a little

9:36bit of

9:37commentary for me, I don't know if that's

9:39true. If you click the learn more link, it

9:41actually tells

9:42you what these checks are doing. So number

9:44of views in a dashboard, there's a specific

9:45bit of

9:46documentation. So it actually gives you a

9:48little bit of context around each of these

9:51things. So

9:53if you're number of views in a dashboard

9:55from the workbook optimizer, blah, blah,

9:57blah, blah,

9:57a dashboard must load all elements before

9:58it can be displayed, the more views a

10:00dashboard

10:00contains, the longer it will take to load,

10:02although it will often involve a redesign.

10:04Reducing the number of views in a dashboard

10:06is often the best way to boost its

10:08efficiency.

10:08So here's a story, you download an

10:10accelerator, you love it, you love

10:12everything in it. But then

10:14you start to have performance issues. And

10:16it turns out to fix the problem, you've got

10:19to redesign it.

10:20So I got asked, what is a template that you

10:23have to redesign? Maybe it's inspiration,

10:27maybe that's

10:27what it should be. But it's just something

10:30you have to be careful of, right? It's

10:32going to be

10:32a bit of work getting these to perform in

10:35the best practice way, you can't just sort

10:37of take these

10:38off the shelf. What I would love is to

10:40tablet to sort of take this and say, look,

10:42okay, what do you

10:43need from this dashboard, click a few tick

10:45boxes, and then dynamically generate the

10:48workbook, that

10:48would be something that this exchange would

10:51be really sort of showcasing that something

10:53unique.

10:54Otherwise, you just have a workbook, a

10:56website with workbooks, and that's tablet

10:58public, it's

10:59nothing new. So that would be something

11:01that again, I think would be a really good

11:03idea for

11:03this. So again, I'll put a link to that

11:05idea in the description below. So check it

11:07out. So that's

11:08one down five. Quite a few times here have

11:12things that need review number of visible

11:15worksheets,

11:16the workbook contains 13 visible worksheets

11:19. Okay, cool. That's good to know number of

11:21hidden

11:22workbook sheets. Wow, there's 114 hidden

11:25sheets. Now this 114 sheets to me that says

11:28, look, we need

11:29to split this up into multiple workbooks.

11:32And you can get into that in another video.

11:34But long story

11:35short, it's got quite a lot of sheets in

11:37here. calculation length, one of these has

11:39591 characters

11:41long. That's crazy. It could be

11:43documentation inside of the calculation. So

11:46active customer

11:47performance value versus reference. But 591

11:49characters isn't that much, especially if

11:51you've

11:51got nested calculations, or you sort of

11:53written out in a very weird way. So you

11:55know, I'm too

11:56shy. Nice to know that that's there. The

11:58data source sales data extract contains 78

12:02level of

12:02detail calculations. That to me just

12:05screams I should have maybe done this back

12:07in the backend.

12:08Okay. When you have 78 LODs, you can

12:11probably push some of that work back into

12:13your ETL pipeline and

12:15sort of streamline your data sources as

12:17much as you possibly can. But all that said

12:20, it did pass

12:21these six things. So number of data sources

12:24, unused data sources, calculation uses

12:26multiple data

12:27sources, all good. Everything is pretty

12:30good on these fronts. So six out of 12. Let

12:33me open up a

12:33notepad here, let's open, let's open a

12:36brand new note. And then we'll get these

12:40going. So let me

12:41just open this up. We'll call this scores.

12:43I'm doing it off screen so you don't see

12:44all my notes.

12:45Now bring it in. And we'll call we'll call

12:48this what's this called a sales expert.

12:50Okay, we call

12:51this and this got a six out of 12. I again,

12:54another feature I would love link in the

12:56description below

12:58is the ability to copy this in some way or

13:01form. This is useful. But I would love to

13:04just take a

13:06PDF of this and put it somewhere and send

13:08it to someone, maybe to a user with all

13:11this good best

13:13practice so they can maybe create a task

13:15list from it even better. What if you could

13:17export a task

13:18list? So you could say okay, here are the

13:20things I'm going to fix. I'm going to tick,

13:22tick, tick,

13:22tick, tick, tick, tick, create PDF outcomes

13:25a PDF with a task list, you can then send

13:27that task list

13:28to someone, maybe you're a reviewer, you

13:30want to send that to someone else. Let's

13:31make this

13:32optimizer actually useful. Let's take

13:33action from this. So I think that would be

13:35really, really cool.

13:36Anyway, enough on this particular one, we

13:38're going to have to go so much faster on

13:40the rest of these.

13:41Otherwise, we could be here for a while. So

13:42we'll do one more from this particular

13:45place. Now this

13:46is strange, because it's not sending me to

13:48actually where I came from. So there was

13:49there must have

13:50been a middle hop there between the pages

13:52to get to where I needed to. So we've done

13:54sales expert.

13:55Let's look at web traffic. Let's go to

13:57completely new world. Let's download this

14:00as well. And

14:01download is complete. We'll put it in the

14:03same folder. And then I'll open up my

14:05folder.

14:06I will open Tableau, we'll just put it in

14:07the background here. And I will get this

14:11back in

14:12and we'll go digital marketing as long as

14:13you just drag it into Tableau. Generally

14:15speaking,

14:16it grabs the file and open it in its own

14:17separate workbook. So just give that a bit

14:19of time.

14:20I can see that it's opening up here, it's

14:21going to open up and it's opening up in

14:23another window,

14:24because when I record videos, I have my

14:27recording window as my secondary screen. So

14:30pop ups and

14:31notifications don't sort of come in the way

14:33. So this is the web traffic dashboard. Let

14:35's go ahead

14:35to server run the optimizer six out of 12

14:39before this diamond eight out of 12. I'm

14:42gonna just

14:43ignore this unused fields option will be

14:45good to be able to tell Tableau Hey, this

14:48particular one

14:49that keeps coming up, just don't check it

14:51again in the future. And then just give

14:54people a tick box.

14:55And then they have to rerun it to get the

14:57items from that tick box if they've ticked

14:59it. But

15:00for this one, I'm just going to ignore it.

15:02So we'll call this nine out of 12. Three

15:04need review. So

15:04a bit better than the previous one number

15:06of views in the dashboard. More than 10

15:08views. Okay,

15:09number of hidden worksheets, 60 hidden

15:11worksheets, quite a lot of worksheets,

15:13calculation length,

15:13Oh, funny enough, exactly the same number

15:16of characters. So it could be a reuse

15:18calculation.

15:19But generally, everything else has gone

15:21pretty well. So let's go ahead and put that

15:23in our little

15:24thing. So we'll call that web traffic. We

15:26'll call eight out of 12. Okay, let's go

15:32back. And we can

15:33go back one more press the hold that let's

15:35go to the accelerator space. One more from

15:38the department

15:39section. Let's go to finance. Finance is a

15:43pretty big one. So let's go ahead and go to

15:46that,

15:46download it and get the score. And we'll

15:49get going. So let's go ahead and open that

15:52I won't

15:53double click it, I will open up the folder,

15:55I will open up Tableau. I'll get the folder

15:58in front,

15:58and we'll get the finance one into Tableau,

16:01that should get us going again. So that

16:04should be

16:04opening, give it a second. And once I've

16:06done this one, I'll close all three and we

16:09'll reset

16:09and go to the next section. So here we are.

16:13And here we go financial statement,

16:16everything's good.

16:17Got a lot of metrics in this one. So this

16:18will have a lot of sheets. Okay, and this

16:20got a seven

16:21out of 12. So let's look at it again,

16:23unused filters going to come up how many

16:25unused fields

16:26are some of these are not that many in this

16:29one. So it's not too bad for need review

16:31the same ones

16:32again. 253 L o DS what is going on here? Oh

16:37, wow, like 253 L o DS definitely,

16:41definitely should go

16:44back to the ETL pipeline. I mean, yeah,

16:46that's just that's just I've got to look at

16:48this. Sorry,

16:50I've got to look at this. Oh, wow. This is

16:52a huge data set. Okay. So first and

16:55foremost, it's

16:56actually good use of folders here. I really

16:58like that the folders are well used here.

17:00But look how

17:01many fields are in each of these. And they

17:04're all used. So it's funny, the list of

17:07unused fields was

17:08quite short, which means this is just a big

17:11data source doing a lot of things. And some

17:15of this

17:15should really be pushed back into ETL. If

17:17we look at this, for example, you've just

17:19got depth, depth

17:21to equity ratio, and you've got the current

17:22month, the previous month, blah, blah, blah

17:24, blah, blah.

17:25And the calculation for this is just going

17:27to be it's going to be dynamic, it's going

17:29to be looking

17:29back and forth based on the specific table

17:31and the context if you're looking at table

17:33calculations.

17:34But if you're doing this amount of

17:36calculations, and you want to keep it

17:39dynamic, fine, do it in

17:40Tableau. But then normally that the way

17:42people asking questions is normally well

17:44defined. So you

17:45can probably bake some of this into the

17:47data set without having to do all of this

17:49stuff, make it

17:50easy for the person to build a dashboard

17:52definitely makes it a lot easier to sort of

17:54track the field.

17:55And all the metadata will be visible to

17:58other tools, not just Tableau, because you

18:00've got to

18:00remember all of these calculations, unless

18:03you're working pretty heavily with the

18:06metadata API and

18:07pulling these out, these calculations only

18:09live in Tableau, they're not available to

18:11you elsewhere.

18:12So maybe do this in Tableau prep, then push

18:15it to Tableau as a data source. And then

18:18other people

18:18don't have to recreate these calculations.

18:20It could also use a published data source

18:22to do

18:23the same thing. But yeah, okay, that's a

18:25pretty interesting workbook there. So if we

18:28go back to

18:29that, that was why is it saying six out of

18:3112? Has it changed? Has it changed? Did I

18:34just score this?

18:35No, I didn't score it got a six out of 12.

18:38So we'll call this finance finance. I'm, I

18:43'm absolutely

18:43adamant something's changed there. And we

18:46'll have a look in the replay, but I think

18:48it went from

18:49seven to six, I think, who knows. But

18:53nonetheless, the workbook loads really,

18:56really fast. What I am

18:57not doing in this analysis is making any

18:58judgment on the design. Okay, I'll come

19:00back to that point

19:01later on. But that nonetheless, that's

19:04everything in the section. So let's go

19:06ahead and close the

19:07three that we've got in here. And we will

19:11get an overview of what's going on with the

19:16section. So

19:17let's go ahead, I'll leave this one open.

19:20And what we'll do is we'll go back to our

19:22score here. So

19:24we'll call this the which section was this,

19:27this was departments. So we did three out

19:31of the six,

19:32and for three out of the six. So 2036

19:37equals 55%. So, you know, how do I know? I

19:46'm being a bit I'm

19:47maybe I'm being I'm being a bit unfair. But

19:49these definitely don't sort of hit the

19:51baseline score.

19:51Now, if you look at some of the reasons

19:53that don't hit the baseline score, those,

19:56you could argue,

19:57those are the bits you have to customize,

19:59those are the bits you have to really make

20:00the template

20:01your own. And but it's interesting. And the

20:04other thing is that the workbook optimizer

20:06is not

20:06testing the kind of things that, you know,

20:08an advanced user would test, it's testing

20:11most of

20:11the things, but there's definitely more to

20:13add to that workbook optimizers. So there's

20:15a good chance

20:16that these scores could get worse over time

20:18, definitely some improvement on the three

20:20that

20:20we've just looked, okay, so that's the

20:22department section. Let's go try some of

20:24the industry

20:25dashboard. So for this one, I'll try to

20:26because I think you're getting the point of

20:28what I'm trying

20:28to get across here. So let's go to wealth

20:31management. This is another probably quite

20:33a popular sector that uses Tableau in lots

20:36of different ways. Sometimes, sometimes

20:38there's an

20:38embedded capability inside of a platform.

20:41So we'll go ahead and save this. Now this

20:43is called

20:44financial services, wealth management. So

20:46they've cut these up into sort of sectors,

20:49if that makes

20:50sense. So let's go ahead, open up the

20:52folder financial services. And let's just

20:54make sure

20:54that's the right one wealth management,

20:57yeah, drop that in. I've got a previous one

20:59open. But we'll

21:00make sure that this new one opens, it opens

21:02up in a new tab. And what I'll do is I'll

21:05close this one,

21:06I'm not going to save it because I now have

21:08my new one on a new screen, wealth

21:10management dashboard

21:12run optimizer. Seven out of 12. Okay, so

21:16let's just go and get that score in. Let's

21:19go to score.

21:21Here we go. So, well, what got a seven out

21:26of 12? What did it fail? Unused field, that

21:32's fine.

21:32Number of using a dashboard, 150 LODs.

21:35There's a pattern to these, there's a

21:36pattern to these,

21:37just sort of starting to build up. And it

21:39makes sense if they were built by the same

21:41group of people with the same sort of

21:42mindset. And it could be also really good

21:44to have a sign

21:45that because these are templates are all

21:47going to have the same sort of problems.

21:48So let's go ahead and test. Let's test a

21:51couple more. So we've tried wealth

21:53management,

21:54point of sales, let's do that. Let's grab

21:56something with a little bit more sort of

22:00fast moving consumer goods focus. Let's

22:03drop that in a straight and tableau. And

22:06then I'll go back

22:07to this and retail point of sales. This is

22:10point of sales is technically till data. So

22:14let's say

22:14you're at the supermarket checking stuff in

22:16point of sales essentially data from this

22:18system. So

22:19let's go ahead and get that in. Okay, so it

22:24crashed. I've just opened tableau again

22:27from

22:27scratch. I'm just going to drop it in. We

22:30'll see what happens again. I've opened up

22:34my activity

22:35model. So just to see if anything starts sp

22:37iking. There we go. That worked this time

22:39around. I can

22:40just blame that on quirks. Sometimes

22:42sometimes these things happen. You are you

22:44ask too much

22:45of your computer and it causes a bug or an

22:48error elsewhere. So let's go ahead and run

22:51the optimizer.

22:52Let's actually select tableau in this

22:53particular case, run the optimizer six out

22:56of 12. So a little

22:57bit lower. Okay, so let's call this POS six

23:01out of 12. So you know, this causes

23:04consistently getting

23:0550% unused fields fine. Number of views in

23:08a dashboard 10 number of usable worksheets

23:1112

23:11number of hidden workbook sheets 101. That

23:15's a lot of sheets in one workbook. Same

23:17number of

23:17characters looks like the same calculation

23:20again, 96 LODs. There's a lot going on.

23:26You know, if I go on this LODs suggestion,

23:28let's see what it actually says.

23:29Competing results of complex calculation

23:32can impact performance. The workbook

23:34contains,

23:35that's just an example error. Often LODs

23:38are used on the fly to address issues of

23:40granularity in

23:41the data source that could be handled prior

23:43to analysis. So exactly what I said earlier

23:45on,

23:45consider tailoring the data source to the

23:48analysis to avoid the needs of many LODs.

23:51So yeah, Tableau tells you right there,

23:53optimizer data source for the question in

23:56hand. So I don't

23:57need to tell you that. Let's go back to the

24:00web page and go back. Any others? I want to

24:03do one

24:04more. I want to do three here before we go

24:06to the next one. I'm going to use this one.

24:08Okay. And for

24:09the last one, I actually want to make a

24:11point about the design, which I haven't

24:12made a point yet.

24:13So just bear with me while I download this,

24:16we test this one, this will be the last one

24:18,

24:18and then we'll go ahead and test and

24:20actually a good looking one as well. So

24:23public set,

24:24citizen service requests. Let's drop that

24:26in and wait for that to open. I won't ask

24:29anything else

24:29of my computer whilst that opens. Give that

24:32a second. So open up on my other screen.

24:34Boom,

24:36there you go. It's loaded up. Let's drag it

24:38onto the screen. There we go. Server, run

24:41optimizer.

24:43I should create a macro to do that seven

24:46out of 12. So again, oh, to fail the unused

24:49fields,

24:49multiple data sources have been have unused

24:52fields, fine, non materialized calculations

24:55.

24:55So this is actually good practice to do. So

24:57essentially, what they're saying is the

24:59data

25:00source public sector CRS citizens has not

25:02been pre computed. What you can do

25:04sometimes in some

25:06data sources, you can go ahead and actually

25:08pre compute calculations. So what that

25:09means is

25:10essentially it's going to go ahead and

25:11calculate them in advance. So if I right

25:13click on this

25:14particular extract, you can see that I can

25:16select compute calculations. Now, if I do

25:18that, it

25:19basically runs the data source and it kind

25:22of has a think and now that that's done

25:24that we can go to

25:25the run optimizer again, and we should get

25:27a better score. There you go. So eight out

25:30of 12, because we

25:31did that fix, it's no longer part of this

25:34one. So that's a very simple free step

25:37there. You typically

25:38do that the end of your workflow,

25:40essentially just goes and pre calculates a

25:42bunch of things that it

25:43can pre calculate and then bakes that into

25:46the part, I'm going to say metadata and the

25:48data

25:49sources kind of two things together. And

25:51then when it's done that runs through the

25:53data source

25:54a lot quicker. So that's a little thing

25:55that maybe you wouldn't have known if you

25:57didn't run the workbook

25:58optimizer. So you had to get another good

26:00thing to do, just run the workbook optim

26:02izer and you might

26:02even learn something. It passes a lot more

26:05this time around. So there's a lot less

26:06than this sort

26:07of orange area. So this is actually a

26:09decent workbook. I'll probably put it in

26:10the as probably

26:12the one the best performing one we've got

26:13so far. Eight out of 12. Let's go ahead and

26:15look at our

26:16scores. It's not it's as good as the web

26:18traffic. So this is the public sector one.

26:20Let's give that

26:22an eight out of 12. Great. And then what we

26:25'll do now is we'll move into the next

26:27section. So this

26:28is the last one. And what I'm going to do

26:32with these is these, I happen to know some

26:35context

26:35behind these. Okay, so and you could

26:38probably argue that because I know the

26:40context with this

26:43section, I'm going to go specifically for

26:45this account engagement one. And I'm going

26:48to make a

26:48specific point about this one. So this

26:50account engagement one is not like the

26:52others. If you

26:53look at it, you can see that it has just a

26:55completely different feel. If I actually go

26:58back

26:58one page, and you just sort of scroll

27:01through, it's very, it's very obvious

27:05whether someone has

27:06spent a certain amount of time looking at

27:08specific things, just by looking at the

27:11workbook. And the

27:13only way I can put it is, it's a bit like

27:16looking at a workman's, you know, workspace

27:19or going to

27:19someone's desk and having a look at how

27:21their desk is laid out. It tells you a lot

27:23about that person,

27:24it tells you a lot about that individual.

27:26And from that you can infer a certain set

27:28of assumptions

27:28about how much care or how little care they

27:31put to their work. Now, I'm not saying that

27:33about these

27:34things. But you can infer that the same

27:37group of people, the same individuals built

27:40some of these,

27:41because I look at all of these, they all

27:44use the same sort of generic color palette,

27:46they even

27:47recycle certain charts and certain styles.

27:50They recycle certain layouts across

27:53multiple

27:53extensions. If you actually download a lot

27:55of these accelerators, you find there's a

27:56lot of commonality

27:58with sections just moved around the

27:59different sort of industry verticals.

28:01Whereas if I come to this

28:03one, this is a very different one. It's

28:06very different fields, much simpler. And

28:08what I will

28:09tell you about this, I bet you this will

28:10get the highest score. I haven't, I haven't

28:12run the test

28:13on this one, actually, I don't know the

28:15score. But my guess is whoever's built this

28:18has put a lot of

28:19thought into it, because there's a bit of

28:21design, there's a bit of layout. And there

28:23's a lot of

28:24simplicity, which actually takes a lot of

28:25work. Because when there's simplicity, you

28:27have to

28:28really think about what's left on the

28:30dashboard to aid the user to get to what

28:32they need to. So

28:33my big bet here is this is going to be the

28:35top scoring one, I'll even go as far as

28:38saying I

28:38think this is going to get a perfect score,

28:41apart from the unused fields. And debacle.

28:44It's also

28:44funny that it's named entirely differently.

28:46So that's, that's how different it is, it's

28:49the only

28:49one with one of these asset names. So it's

28:51obviously come out of a system, and they

28:53haven't

28:54just bothered to rename it. But nonetheless

28:56, you can really, really tell when something

28:58's,

28:58you know, has a different contributor sort

29:01of working on it. So let's go ahead and

29:03open this,

29:04I'm really curious what this gets. So my,

29:06my guess is it's going to get 11 out of 12.

29:08That's my sort

29:09of hunch. So let's wait for this to open up

29:11. So open up another screen, here we are

29:14full screen,

29:15drumroll, run optimizer, 11 out of 12. Like

29:19I said, and yeah, this this just proves my

29:24point,

29:24you can you can just tell when someone has

29:27spent time working on something. And whilst

29:30I haven't

29:30been looking at the designs previously, it

29:33's no surprise that this well designed work

29:36book,

29:37the person who's built it has also gone and

29:39spent a good amount of time looking at best

29:41practice,

29:41and has actually implemented it. And it's

29:44funny, because earlier on, I said, well, in

29:47a template,

29:47you might leave a lot of data sources and

29:50fields in there, so the user can customize

29:52them. But what

29:53this says about this one is actually the

29:55person who's built this has really focused

29:58this dashboard

29:58on the question in hand, account engagement

30:01, okay, so it's giving you a high level

30:03overview,

30:04it's giving you a starting point. And it's

30:06got a bunch of interactivity, click on an

30:08account to

30:09filter it. So if I go click on a singer,

30:11singer, lumber, it filters the whole

30:13dashboard, it becomes

30:14a bit of a story. And I can then go into

30:16some of these and start to enhance these.

30:19And there's only

30:20one tab, there's not 100 tabs, there's just

30:23the one tab. And if I go look at the data

30:25source,

30:25it's look at look how focused this data

30:28source is, okay, like, it's just so simple,

30:31and they've just

30:31gotten rid of everything else, because it

30:34just only uses what it's supposed to use.

30:37Now, I want

30:38to see if I can show the hidden fields. And

30:39there you go. So this is what you would

30:41have got. If

30:43you'd not touched it, okay, so this is the

30:45all the fields you would have got the gray

30:46ones have

30:47basically been removed or hidden,

30:49essentially. And some of these are also

30:50actions. But let's just

30:52ignore that for now. And if I untick that

30:54box, you can see that look, only the ones

30:56that have been

30:57used are actually left in there. So this is

30:59something you can do, hide all unused

31:01fields,

31:01once you're done building it, and you're

31:03certain that everything's there. But you

31:05can show them

31:05in case you need to right click and unhide

31:07them and bring them back. So that's a

31:09really

31:09important bit of best practice. So can you

31:12imagine that got an 11 out of 12? Okay, so

31:16for my overall verdict on these things, I

31:20think I think these have actually scored

31:24better than I

31:24thought they would. And that's again, not

31:26looking at design, not looking at

31:27functionality, I'm just

31:29looking at the workbook optimizer and what

31:31it thinks the workbooks worksbooks could do

31:34better

31:34to perform better. And it's telling if I go

31:37back to the accelerators tab here in Table

31:40au Exchange,

31:41you can see there's a whole ton of acceler

31:44ators in here. And I also wondered why they

31:48went with

31:49icons for a lot of these. And I part of me

31:53wonders if they went with icons, because

31:56they know that

31:57the thumbnails, the dashboards themselves

31:59aren't strong. I don't know. I don't know.

32:01I don't know.

32:01Is it me? Do you think that's the same

32:03thing? Am I just making that up? If you're

32:05actually someone

32:06who works on this, let me know. But I think

32:09if you're going to show people templates,

32:13why not do what you do on Tableau public,

32:15which is lead with the thumbnail, why not

32:17do what you

32:18do with Tableau server, Tableau online,

32:20which is now Tableau cloud, why not do what

32:23you do there

32:24and lead with the thumbnails. My hunch is

32:27that the thumbnails all looked very samey.

32:30And so they

32:30decided to go with icons because they felt

32:32they were more accessible and easy to look

32:35at. But I

32:35think if they were well designed and

32:37actually simple for people to use, people

32:39would be able to

32:40work with them. The other argument is you

32:42could say that maybe people have a stronger

32:44affinity with

32:45icons from tools they use. So for example,

32:47if I go over here to the top and I search

32:49Google

32:50and it goes and finds everything around

32:53Google. Well, I know that's Google

32:55analytics. I know

32:56that's Google ads. And you can sort of

32:59start to get a good sort of feel for stuff.

33:01Whereas if I

33:02just had dashboards, you could argue that

33:03that was sort of harder. Maybe that is the

33:05more likely

33:06reason. But if I go to the accelerators in

33:08general and I'll just scroll down and stop

33:11and click on

33:13one scrap cost based. I can tell you that

33:16this one is probably going to have a score

33:19of six out of 12.

33:21So let's go have a look. Download. Okay, it

33:23's opened up on another screen. Let's drag

33:26it in.

33:26Drum roll. I think it's going to get a six

33:28out of 12. Let's see what it gets.

33:30Nine out of 12. So I'm wrong. So that that

33:35theory is just completely out the window.

33:36You can't tell

33:37the score of a word, but just by looking at

33:40it, that's complete nonsense. But I still

33:43do think

33:43what I said before about just by looking at

33:46a word, but you can tell what kind of

33:49person has

33:50built it. I can't stress that enough. If

33:52you go to Tableau Public, you'll get a

33:55similar sense

33:56of where a user is in their learning

33:57journey just by looking at the dashboard.

34:00That is the end of

34:01the video. So all of these scored between,

34:05well, let's say on average, let's say like

34:09seven out of

34:1012. If I'm being sort of reasonable, seven

34:12out of 12 was the average scored out.

34:14Everything in the

34:15accelerators. If I was being a little bit

34:18more generous, I'd actually say seven out

34:21of 11 because

34:22every single workbook failed the unused

34:24field test. Not a single one passed it. So

34:27in my opinion,

34:28you shouldn't be checking it because it's

34:30actually quite a hard one to pass unless

34:32you're at the very,

34:33very, very final stage of going to

34:34production and you want to hide everything.

34:36So yeah, what do you

34:40think? What do you think of the workbook

34:41optimizer? What do you think of acceler

34:43ators? What I really

34:45like is that the community is now starting

34:46to build more accelerators. So you'll get

34:48to see,

34:49I think, a better sample of what workbooks

34:51can be. And hopefully these come with Table

34:53au Cloud

34:54and Tableau Server as well. So more of the

34:56goodness that the community has already

34:58been

34:58able to share in the past through Tableau

35:00Public makes its way onto people's everyday

35:03work setup.

35:06So yeah, thanks for watching and I'll catch

35:07you in the next video.

35:08[END]

35:11[ Silence ]