How Tableau Pulse Works | New in Tableau 2024.1

0:00Hey, it's Tim here. In today's video, I'm

0:01going to show you how Tableau Pulse works.

0:04As ever, let's get stuck in. So we'll start

0:06off right from the back by saying that

0:08Tableau Pulse has just gone into production

0:11as it were. It came out of beta and it was

0:13launched in

0:1424.1. So I'm recording this essentially

0:17almost a week after it's been made public.

0:20I did go

0:20through a long video going through my first

0:22impressions of how it works. If you want a

0:24more

0:24rough cut of how that experience was, go

0:27ahead and check out that video. This video

0:29is going to try

0:29and synthesize everything you need to

0:31understand how Tableau Pulse works. So I'm

0:34going to pick up

0:34from where I left off in my last video

0:36where I showed you just how to enable Table

0:38au Pulse.

0:39Today, we're going to get straight into the

0:41action by going straight into Tableau Pulse

0:43.

0:43As a refresher, you'll see that once you've

0:45enabled Tableau Pulse, you'll see that it

0:47turns

0:47up here on the left-hand side. And when we

0:49go ahead and click on that, it takes us

0:51through to

0:51the experience. So we'll go ahead and click

0:53that and you'll see that we land inside of

0:55Tableau

0:55Pulse. Now there's a couple of things to

0:57bear in mind. I don't have any available

1:00metrics and

1:00there's nothing for me to browse. So this

1:02shouldn't be your first experience if you

1:03're using Tableau

1:04Pulse because in essence, you should

1:07already have metrics available to you. So

1:09the experience I'm

1:10going through today is essentially how to

1:13build those metrics. In another video,

1:15separate video,

1:16I'll go through a more detailed, simpler

1:18video on how to use the metrics if that

1:20makes sense.

1:21So let's just focus today on how to build

1:23the metrics and the metric definitions. And

1:25then

1:26the next video, we'll talk about how to use

1:28them. Okay, so we're here, we've landed in

1:30Tableau

1:31Pulse. I'm a creator, I'm an author, I need

1:33to build metrics for people to use. The

1:35first thing

1:35you're going to notice is there's a nice

1:37big blue call out over here on the right-

1:38hand side to

1:39create a new metric definition. And this is

1:41actually quite important because Tableau

1:43Pulse

1:44breaks down the creation of metrics into

1:46two steps. The first one is the creation of

1:49a metric

1:50definition and the second is the creation

1:52of a metric. And those two kind of happen

1:54in one go.

1:55If I just open up my sidebar here and I go

1:57to this documentation, I'll put a link to

1:59this in

1:59the description. Tableau have actually done

2:01a really good job of detailing this split

2:03here in

2:04this diagram. I actually really like this

2:05diagram. It's a good way to conceptualize

2:07the creation of

2:08metrics, whether or not you're building for

2:10Tableau Pulse or a dashboard, this is a

2:11really

2:12good way of thinking about it. So if you

2:14look at this example, the metric definition

2:16essentially

2:17needs three things. It needs a measure,

2:19some sort of time dimension and the data

2:21source essentially.

2:22And then from that, we're able to add

2:24additional context. So once you have those

2:26three pieces of

2:27information, the additional context,

2:29essentially the step two is defined as a

2:31metric. And for the

2:32metrics, we start to drill into sort of

2:34subsets of that specific definition. So if

2:37you can see here

2:38that down in this metric, we're doing year-

2:40to-date for technology and a couple of other

2:42categories,

2:43and it's essentially still on sales. And on

2:45this one on the right-hand side, we're

2:47still doing

2:48year-to-date, but instead we're doing it

2:49for office supplies. So that is classed as

2:51a separate metric,

2:52but they're all derived from the same

2:54definition. And that's a super important

2:56thing to understand.

2:57So when you go ahead and build for the

2:59first time, think about your definition,

3:01and then you can

3:02enhance that with various metrics that feed

3:04off that original definition. You'll see

3:06this later on

3:07in the video, I'll call it out. But if we

3:09head back to Tableau Pulse, when we go

3:11ahead and hit

3:11the new metric definition, you'll see we

3:13get a list of data sources. Now, in my

3:15previous video,

3:16I highlighted that the data sources you see

3:18here are from your Tableau Cloud instance.

3:21Don't forget

3:22Tableau Pulse is not available for Tableau

3:24Server, so you can't get these from Tableau

3:27Server.

3:27They're only from Tableau Cloud. And I've

3:29actually got a data source which has my

3:31YouTube analytics

3:32that essentially been piped into Snowflake

3:35from Frivetran through the YouTube API. So

3:37we'll go

3:37ahead and click on that and hit connect.

3:39And once we hit connect, we get straight

3:41into the definition

3:42building experience. Let me close my

3:44sidebar so you get a bit more space to see

3:45this. Now,

3:47let me just call out a few things. When you

3:49land in this space, there's two parts of

3:51the journey.

3:52There's the definition, which is what we're

3:54going to do first, and then there is the

3:55insights. The

3:56insights are more of a behavior of how the

3:58metric works with the metric definition.

4:00And so we'll

4:01come to the insights a little later once we

4:03've built out this metric definition a

4:04little bit

4:05more. The next thing you'll see if I just

4:07switch to red here from a highlighter is

4:09the data source.

4:09The data source is essentially what you've

4:11just connected to. And what I will say is

4:13that it's

4:13important to really pay attention to

4:15metadata, I think, in this product because

4:18that information

4:19is used throughout the product to do

4:20different things. So I'm going to call this

4:22subscribers.

4:23And we're going to say that this metric

4:27definition tracks subscribers. Again,

4:29notice my language. I'm

4:31calling this the metric definition rather

4:34than just the metric. So tracks subscriber

4:37change

4:37month to month linked to videos. That's

4:40exactly what this metric does. Sort of a

4:44weird way of the

4:45way YouTube API works. It actually pins

4:47most subscribers to a specific video unless

4:49you just

4:50subscribe out of your own free will. Let's

4:52go ahead and hit full stop on that. And we

4:54've defined

4:55the metric definition. We've defined the

4:57description. Let's have a look at the

4:59measures

4:59here. And so here when we click on the drop

5:01down, you can see it's pulling from the data

5:03source.

5:04And the thing I'm interested in here is

5:05subscribers. So as soon as I start typing,

5:07you can see it searches this. Now for this

5:10particular metric, there's a problem. I

5:12have

5:13subscribers gained. I have subscribers lost

5:15. But what I don't have is a net subscriber

5:17count. So

5:18let me do two things. Let's assume that's

5:20not a problem. Let's just hit subscribers

5:22gained.

5:22And you'll see that it automatically fills

5:25the aggregation and the type of maps that's

5:27going to

5:27work. So you see here running total works

5:30out of the gate and you get non-cumulative

5:32as well there.

5:33And so we can select one of these and you

5:34can start to sort of play around with this.

5:36Now

5:37the key thing you're missing is a time

5:39dimension. Don't forget we needed a measure

5:42, a time dimension,

5:43and then a data source. We've got a data

5:44source. We've got a measure. The last thing

5:46is a time

5:47dimension. So for this, this is actually

5:50linked to the date that the data is being

5:52collected on. So

5:53if I select that date, as soon as we give

5:54it all three pieces of information, we

5:56immediately get a

5:57metric applied. Now again, notice I said a

6:00metric because what is strictly happening

6:04here is that

6:04this on the left hand side is your

6:06definition and this on your right hand side

6:09is actually a metric.

6:11It's just lacking some sort of specificity,

6:13if that makes sense. The specificity we

6:17have at the

6:17moment is month to date. You could choose

6:19another time context, but month to date is

6:22the default

6:22context. So it's sort of an interesting one

6:25because you have this sort of unified

6:28interface,

6:28but you have to really remember that on the

6:31left hand side you're dealing with the

6:33definition,

6:33on the right you're dealing with the metric

6:35, and those two are not the same. I'll

6:37explain why that

6:38is the case later in this video, so please

6:40bear with me and you'll see why this is an

6:42important

6:42thing to understand. Okay, so now that we

6:45've done this, we can go down and you can

6:47see you get a few

6:48other options and you can kind of get the

6:50gist of how this works. It's actually quite

6:52easy to set up

6:52the metric. That is it, that's done. I mean

6:54, I didn't even sort of celebrate that that

6:56's how easy

6:57it was and it really was that easy and I'm

6:59not just a fanboy here, just a championing

7:02tableau.

7:03This is really a very accessible way of

7:05creating metrics in a quick digestible

7:08manner. Now,

7:10remember earlier on we said we had a

7:11problem and this was with this metric,

7:13subscribers gained

7:14and subscribers lost. What I'd actually

7:16like to do is to create a new metric here

7:18called

7:18net subscribers, but you'll notice there's

7:20no interface for me to do a calculation

7:22just yet. So

7:23in this value area here, if you scroll down

7:27, because this right hand space is the

7:30definition,

7:31one of the things you can do is you can

7:33create what's called an advanced definition

7:35. You can

7:35see that right here, okay, and if we go

7:38ahead and click on that, what it does is it

7:40takes us to

7:42something very familiar to desktop creators

7:44and that is the ability to just go into

7:46tableau and

7:47build your metric. You'll notice that my

7:49data source is actually a data model across

7:50two data

7:51sets and I can go and access every single

7:53field as I would do and the way this works

7:56is very similar

7:57to the experience you get with published

7:58data sources in web authoring. You're

8:00essentially

8:00able to go ahead and create your published

8:02data source. You can kind of model what you

8:04're looking

8:04at here on the right hand side in the

8:06canvas, but it never creates those charts.

8:07It's just sort of

8:09an area where you can model things. So what

8:11we can do here, we can create a calculation

8:13and we can say sum of subscribers gained

8:19plus, sorry, minus sum of subscribers lost

8:25and now that we've done that, we can call

8:28this net subscribers, okay, and we'll give

8:33that a bit of

8:34a space there. Net subscribers, that is our

8:35metric, that is a much better way of doing

8:37this. We'll hit

8:38apply and hit okay and you'll notice

8:40nothing has changed because we've just done

8:43what we do in

8:44desktop. We've created a metric down here,

8:47okay. Now to build that metric, notice the

8:50same three

8:50things that we needed before are also

8:52present. We already got the data source,

8:53that's part of

8:54the definition. What we now need to add to

8:56this view is the measure and the time

8:58dimension. So go

8:59ahead and let's drag the net subscribers to

9:02the measure and you'll see that we

9:04immediately get a

9:05bar chart telling us something, so 3000 and

9:08if this is all we did, the metric would

9:10just track

9:10this number. This is not a good reflection

9:13of my actual subscriber count because the

9:15data source

9:16only goes back a bit of a few months

9:18because of the way that 5tran and YouTube

9:20API work.

9:21So let's go ahead and grab a date and the

9:24date we want here is date today and just

9:27like you can do

9:28in Tableau, you can choose the context for

9:30that date. So rather than the year, I'm

9:32actually going

9:32to go back down here and choose a

9:34continuous day if that makes sense. So we

9:37get a day by day blow

9:38of that and there you go, that's something

9:40a little bit more familiar. Now this chart

9:42is a

9:43representation of what's driving the pulse

9:45metric but it's not actually the pulse

9:47metric,

9:48it's just a visual way for you to see some

9:50context. So what do we do from here? Well,

9:54we can go ahead and hit apply and before I

9:56actually do that, notice that you can sort

10:00of

10:00do some filtering on this metric in here

10:03and this is not the same as letting users

10:06choose the

10:06filtering if that makes sense. So a good

10:08way to think of it is if I was building

10:10this advanced

10:11metric and what I wanted to do is exclude

10:14certain videos or exclude certain subsets

10:17of my data from

10:18being part of this metric, I could actually

10:20do that here in the advanced element. So I

10:22can go

10:23ahead and get the video title. Let's say

10:27that I didn't want a specific video to

10:29contribute

10:30to this. Let's say as I go to what is Table

10:33au, the reason I might exclude this is

10:36because

10:37this is the biggest video on my channel so

10:40it has a rather large skew on what's going

10:43on. So if I hit

10:44okay and we say exclude and we hit apply,

10:46you'll see in the background the metric

10:48changes but the

10:49biggest change was actually in the axis and

10:51you can see that that's now taken into

10:52effect and as

10:53soon as I remove it my metric sort of adapt

10:55s to adjust for that. It's a very slight

10:58change but

10:58in the more recent months it hasn't had a

11:00bigger impact, in the previous months it's

11:02like 50% of my

11:03channel grade. So we'll just leave this at

11:05this and we'll hit apply. Now when you hit

11:08apply you

11:08get a reminder about your selections. This

11:11is because if you don't do all these three

11:13things

11:13you'll get an error so this is like a nice

11:15way of reminding you that this is what you

11:17need. We have

11:17all of these, we have a measure, we have a

11:19time dimension, we don't need filters, it's

11:21optional,

11:22so we'll hit apply and that is how you get

11:24around that problem of being able to create

11:27a metric from

11:28a calculation inside of this window and you

11:30'll see that this is now a little bit better

11:32, it's a

11:32little bit clearer and we get a better view

11:34of what's going on. Now one thing I didn't

11:36touch on

11:37in that step is that we have a bit of an

11:39issue with running totals inside of that.

11:42If I go to

11:43replace advanced definition and instead of

11:45the standard measure I've got here I'll go

11:48and do

11:48running total, you'll see that it generates

11:51a running total fine but if I hit apply and

11:53I hit

11:53apply again I get a bug and I don't know if

11:55this is an issue just because it's out of

11:57the gate but

11:58I've had this sort of bug consistently

12:00since I've been using Tableau Pulse and

12:02admittedly it's only

12:03been public for one week so I'm sure this

12:05will get fixed but nonetheless I just

12:07wanted to call that

12:08out if you're having that experience. Okay

12:10so let's go ahead and undo that change, let

12:12's go

12:12ahead and remove the running total and hit

12:15apply and now if we hit apply again you'll

12:17see it

12:17magically works there's no problem now and

12:20we get back. So that is our advanced

12:22definition now.

12:23The way the measure works you kind of get

12:25some nasty sort of Tableau stuff in the

12:27background,

12:27you'll see there that it doesn't call it

12:29what I called it which was net subscribers

12:31it calls it

12:32calculation 105 whatever so that's always

12:34something that happens sometimes if you

12:36look in the Tableau

12:36Workbook XML you can see that kind of

12:38coming through here in the product but

12:40nonetheless

12:41just be aware that these are just sort of I

12:43guess sort of rough edges that will get

12:44smoothed out

12:45over time but nonetheless the metric is

12:47working fine. The second part of this is

12:50then telling

12:51Tableau Pulse how to interpret those

12:53insights and this is the easiest part to

12:55miss I believe because

12:56you can go ahead and hit save definition

12:58but I think you have to get into the habit

13:00of going

13:00into the second tab because if you go into

13:02the second tab you'll notice that there are

13:05no

13:05insights to preview right now what's

13:06happening is in the background Tableau is

13:08already processing

13:09this information so we'll go ahead and it

13:11will do this quite quickly and we'll be

13:13able to come back

13:14to this and look in the future but you can

13:16tell Tableau here what happens when the

13:18value is going

13:19up. Obviously what happens in the other

13:21direction is implied so we don't need to

13:23always specify that

13:24but nonetheless you can tell if going up is

13:27favorable which it is in this case and you

13:30can

13:30also change the insight type so you get

13:32these several insight types here so you've

13:35got five

13:36six insight types and you get a little bit

13:38of a description of what's going on and the

13:41different

13:41types of insights that you can get and you

13:43can disable certain ones if that insight

13:45type doesn't

13:46work you can disable them you can just

13:48change them and switch them off so you do

13:50have a bit of

13:50control into that. Now if we go back into

13:53the definition what I want to do is finish

13:56building

13:57this metric because once we finish defining

14:01the definition everything else defines

14:05additional

14:06metrics okay so it's not super clear to me

14:09but I sort of figured this out over time

14:12these options

14:13down here kind of drive that behavior so

14:15the number format is just a general sort of

14:18description

14:19of the number and if you click on this you

14:21do get the ability to change to currency or

14:23percentage

14:24and you can you can kind of use a

14:26description so we can say unit expression

14:30singular so you can say

14:31subscriber and we can say here subscribers

14:36and that is it and now you'll notice that

14:40nothing

14:41nothing really changes here nothing really

14:43changes in the definition I think this will

14:44have a sort of

14:45a big play further down the line that this

14:48bit here this is where I think stuff really

14:51matters

14:51because the dimensions you add here end up

14:54being what end users can use to customize

14:57their metrics

14:58essentially so the definition you give has

15:00to include these dimensions otherwise users

15:03can't

15:04really build segmented metrics for

15:05themselves for example if I was using super

15:08store I couldn't use

15:09the categories to generate a metric for

15:12each category so in my case I will use a

15:15video title

15:16because I want to be able to create metrics

15:18for a specific video I might also use other

15:21information

15:22about the channel so I might say as a very

15:25simple example the channel title if I was

15:29tracking

15:29multiple titles I might also use let's say

15:32tags and tags are something that I think we

15:36do get in

15:36here I don't know why I'm scrolling down I

15:38can just go I can just go and search these

15:41things

15:41and then the final thing I might want to do

15:44I think that's honestly enough actually

15:47those those

15:48those are enough to show you what's

15:49happening and you can see that they get

15:51added here at the top so

15:52you can kind of think of everything that I

15:54'm sort of circling in this blue box as

15:56being part of your

15:58metric definition and your metric control

16:01all in one package and as soon as you're

16:03done we can go

16:04ahead and hit save let's hit save

16:06definition and here we are we've built our

16:09metric definition and

16:10the first thing you get is actually a

16:12metric out of the get and you can see it

16:14immediately has

16:15insights and I think that's a super

16:17powerful thing here I've just added the

16:18data source and the value

16:20comes out of the system straight away I can

16:22immediately start working with this I don't

16:24know

16:24if it processes it in the background and in

16:26the time I've taken to sort of talk to you

16:28it's done

16:29it but nonetheless um it's here now I will

16:32use this opportunity to just send a bit of

16:34feedback

16:35to tableau this back button it's it's wild

16:38because in this particular instance it

16:42sends you back to

16:43the metric definition it's always

16:45contextual to where you've just come from

16:48which means depending

16:49on where you are across the whole of table

16:51au pulse that back button can mean different

16:53things

16:54I initially thought this back button was

16:56going to go to tableau pulse the landing

16:58page so I could

16:59see all my other metrics but you'll see

17:00that if I hit back it goes back to the

17:02metric definition

17:03and it asks me to select the data source

17:06which is which is slightly odd I just didn

17:08't expect it to

17:09go right to the beginning of the previous

17:11experience if anything I would expect it to

17:13go

17:13back to the editing experience of the

17:15metric I've just created nonetheless um

17:17this is slightly odd

17:18and it can leave people confused and

17:20thinking hey what happened and when you hit

17:22close you go back

17:23to tableau pulse and you're landing on this

17:25page and you're like hey where's the metric

17:27I just

17:28created it's not here so here's the thing

17:30the metrics you've just created you've not

17:34yet had

17:34a chance to follow said metrics so if we go

17:37to browse metrics you'll see that that

17:39metric is

17:40indeed there if I click on it I do actually

17:42get to see this and remember earlier on

17:44when I said that

17:45you create your definition and your first

17:48definition you get a free metric with it

17:50essentially that first metric is the same

17:52as the definition without any segmentation

17:55so

17:55what does that mean well let's go to this

17:58metric and let's add some segmentation and

18:00you'll see

18:01how you can save additional metrics having

18:03defined them so we'll go ahead to this we

18:05'll hit adjust

18:06and then we can go ahead and start

18:08searching for specific things so if I want

18:10to find all my videos

18:11that say tableau pulse um you'll see that

18:14it's able to go through and we can go ahead

18:18and just

18:18tick all of these um and we'll hit okay

18:21these are the five so those five videos are

18:23specifically

18:24about tableau pulse and they're all from my

18:26channel there's nothing special there

18:28unfortunately um I'll come back to this

18:30point later but if you've not done your

18:32data prep it

18:33will show here so the tags are not in the

18:35format I was expecting I was expecting them

18:38to be sort

18:39of separate items but they're all just sort

18:41of everything I put into that space in

18:43youtube so

18:43it's not perfect but nonetheless we won't

18:46use that um I can choose different contexts

18:48of time as well

18:49this is super important depending on the

18:51context you have you can choose week to day

18:54and you know

18:54this changes the metrics so as soon as I

18:57hit this um tick what that does is it saves

19:01an instance of

19:02a metric and you'll see in the last week

19:04nothing has happened and a lot of you will

19:05see this the

19:06first time you use tableau pulse because

19:08your data has to be running to today's date

19:11if you look at

19:12this tableau pass is always basing itself

19:14on today and if your data doesn't have

19:16anything today this

19:19is what you'll see however if I go back and

19:21I adjust this metric let's say we don't

19:23want to do

19:24week to date let's do month today and we

19:26hit plus then you'll see that we do get

19:28some information

19:29back and it stops right here okay and this

19:32is this is super useful because um I spent

19:37about five

19:37minutes for the first time just trying to

19:39understand what was going on and I couldn't

19:41figure it out and

19:42I realized oh it's because there's no so

19:44there's no there's no data in the context

19:46that I have so

19:47here we have the information these are just

19:50the videos related to tableau pass so over

19:53the last

19:53month and there are about 22 subscribers

19:56who've come from videos relating to tableau

19:58pass

19:59now we've done this a couple of times we've

20:01created variations of this metrics a couple

20:03of times so let's go back and again when

20:05you hit back this time it takes you

20:08somewhere different

20:08to where I'd expect notice that rogue

20:10instance where we created a metric that

20:12didn't have

20:13anything in it that was instantiated

20:15immediately you didn't have really an

20:17option to save it or

20:18delete it and that's been created and the

20:20most recent one we created is also here so

20:2222 here as

20:23well so there's a little bit of house

20:25keeping you have to do if you're just busy

20:26clicking these

20:27metrics and not sort of paying attention

20:29you can come back here and find you've got

20:31three or four

20:31metrics you didn't intend to have so you

20:33can obviously manage those by just going

20:35here

20:35seeing the details and as soon as you go to

20:38the details you can go back in there and

20:40actually

20:41go ahead I thought you could delete them

20:43here but actually that's wrong um what am I

20:46what am I

20:46missing here um okay I thought you could

20:49delete it here maybe I'm missing something

20:52I could swear

20:53I found a way to delete a metric but maybe

20:55I was deleting metric definitions um so um

20:59maybe this

21:00will come back to me but I hope I hope you

21:02can delete these rogue ones because you saw

21:05how easy

21:06it was for me to create this null metric

21:07and they're kind of frustrating as well

21:09because they

21:09don't need to get experience and if they're

21:11discoverable then hey you don't want people

21:12to be seeing these but nevertheless um this

21:15is this is sort of the experience now let

21:17me go

21:18ahead and hit follow on these two and if I

21:20go back to the Tableau pass landing page by

21:22hitting that

21:24you'll see that those two metrics I created

21:26now land here on my landing page so we've

21:29got these

21:29nice metrics available to us here and we

21:32can kind of just get using them so if I go

21:34back into the

21:35subscribers this is it this is how it's

21:37working now at any point if you want to

21:39change this if

21:40you want to play around with this if you

21:42click on these three dots at the top you

21:43always get the

21:44option to edit the definition and that

21:46takes you back to this experience and you

21:48can obviously go

21:49back and change the insights and when you

21:51do this it changes the context each time

21:53this is also a

21:54good place to go back to this insights tab

21:55and show you that now it's generated some

21:58context and

21:59what these really lend themselves to are

22:01the different types of insights the Tableau

22:02pass can

22:03generate we've already been going for 20

22:05minutes here so I don't want to make this

22:06video longer by

22:07going into the different types of insights

22:09you can get I'll do this in a separate

22:11video but just be

22:12aware that there's a broad range of insight

22:14types and they generally map to these ones

22:16that you can

22:16see here but in terms of how they play out

22:19and it probably helps to be able to see how

22:21those

22:22interactions work so I'll come back to that

22:24a little later the other thing I'll also

22:26highlight

22:26as I sort of scroll up and down here is

22:28that look if you've not done your data prep

22:30up front this is

22:31just still not going to give you nice

22:32results you can see here that the tags the

22:34tags are just

22:35causing so much noise in here and there's

22:37valuable information because I use these

22:39tags on videos

22:40where it makes sense but there's just they

22:42're just not being cleansed properly so what

22:44I would love

22:44to do is to have a way of being able to use

22:47this information and just pass this out so

22:50Tableau can

22:50understand that although these are lists

22:52actually all we care about is the items in

22:54each of these

22:56in each of these sort of list of values

22:58arrays of values as it were if you're a

23:00programmer so

23:01it's just a little bit something sort of to

23:03bear in mind if you're if you're not doing

23:05your data prep

23:06up front you're gonna have to do it in

23:08Tableau prep Tableau desktop in your data

23:10source or do

23:11it back further back somewhere more

23:13observable in something like dbt or in a

23:15data engineering

23:16operation that you're that you're running

23:18okay the other thing I have to say about

23:21this if I just go

23:21back and save this we'll just save this as

23:24is we haven't really changed anything is I

23:26would like

23:27the ability to be able to put what I would

23:29call the wild card in here that updates as

23:31a data update

23:32so at the moment you can see the video

23:35title here I can go ahead and type you know

23:37Tableau desktop

23:39and all the videos related to Tableau

23:41desktop come up now the problem is is you

23:43know when I when I

23:46deselect one it deselects and if I hit

23:49exclude it excludes all of them so if I do

23:52exclude and then

23:53I just sort of delete this you see that it

23:55says that all of them have been excluded so

23:58there's no way for me to do like a wildcard

24:01search on the data set if that makes sense

24:03there's no way

24:04for me to say anything that mentions this

24:06specific term always added to this metric

24:09because then I

24:10think that makes metrics a little bit more

24:12meaningful and it makes these drop downs

24:14actually represent what we can do already

24:15inside of Tableau today so that would be

24:17one additional thing I'd love to see in the

24:19future but nonetheless I'm not going to

24:21create a new

24:21metric I don't want to create more noise

24:23right the very final thing I know we've

24:25been going for some

24:26time we're nearly 30 minutes in is at this

24:28section here this is where really the magic

24:30starts to

24:31happen if we spend a bit of time just

24:33working with this metric you can see that

24:35obviously we've got

24:36the overview and the breakdown in here and

24:38if we go to the breakdown we actually get a

24:40little bit

24:41more of a detailed breakdown and you can

24:42see this is a genuine reflection of my

24:44channel the Tableau

24:45desktop crash course is driving a good

24:47number of subscribers each month and if we

24:50sort of go down

24:51the list other videos have a slightly

24:53smaller impact but nonetheless this is very

24:56representative

24:57of what I know and what works if I select

24:59video tags though you can see the tags

25:01again are a bit

25:02noisy and this makes us basically pointless

25:04because a large bunch don't have a value

25:06and the rest are

25:07just I mean what can I really do with that

25:09okay there's not much it's too much

25:11information

25:11I'll get back to the overview you can

25:13obviously hover over this and work with it

25:16which is quite

25:16nice and I like the fact that this is

25:18contextualized I do need to get my data

25:20source to update a little bit so there's a

25:22little bit of an issue with my sort of data

25:23pipeline in

25:24the background but here's the next really

25:27sort of exciting things this section here

25:30is where the AI

25:32comes in so they call it GAI so generative

25:36AI right these questions are part of a sort

25:39of a

25:40system that Tableau's built where

25:42essentially the way that the platform works

25:45is the insight is

25:46actually generated from your raw data that

25:48is just just think of it as maths being

25:51done on your data

25:52and essentially those are facts those facts

25:54are then presented in order of preference

25:56in order of

25:57sort of what's working and what's not and

26:00then once that's done it's summarized by

26:02generative AI

26:04and then what you get back is essentially

26:06these these sentences and these sentences

26:08are a good way

26:09of helping the user go to the next question

26:12understand what's going on and so the

26:14computations

26:15that drive these questions have already

26:16been done in the background but what these

26:18questions are is

26:19sort of tempting you to explore those

26:21journeys take those journeys with your data

26:24so if we go

26:24ahead and say okay which video title

26:26decreased the most which video title had

26:28the most negative impact

26:30on my subscriber count you can see here

26:31that it's actually what is Tableau and this

26:33makes sense

26:34because the way you have to think of this

26:36metric is that the metric isn't about which

26:39video lost me

26:40subscribers remember our metric was

26:42actually net subscribers it's not

26:45subscribers lost it's

26:46subscribe it's not subscribers gained it's

26:48net subscribers so I could go ahead and

26:50recreate this

26:51with subscribers lost and this sentence

26:53would make more sense but all it's saying

26:55is that I got 19

26:57less subscribers this month compared to

26:59last month for this particular video okay

27:01and it gives us a

27:02nice sentence that actually breaks that

27:04down a little bit and remember all the

27:05terminology we

27:06did before for singular versus plural that

27:09's all in here that's all being used to kind

27:11of make this

27:11work really really well and as we click on

27:14these uh notice every time I click on

27:16something here

27:17it gets added just below here okay so let's

27:20go ahead and see which video title had high

27:22subscribers

27:23so let's go ahead and hit that again you

27:25get the Tableau desktop crash course you'd

27:26already seen

27:27that and as you click on these the metrics

27:29sort of go down so I can kind of just click

27:32on these

27:32and you'll see that all the metrics sort of

27:34come through and remember some of these

27:36were some of

27:36the noisy ones we were seeing on the

27:38insights tab before so that was already a

27:40preview of what you

27:41were going to see and it's essentially just

27:43enabling and disabling them behind the

27:45scenes so

27:46it's a pretty good platform and you know my

27:48data is pretty simple I'm just tracking one

27:50thing

27:50subscribers but in Superstore and in other

27:53data sets you might have slightly richer

27:55context that

27:56you can work with one thing I will say if

27:57you're working with Superstore make sure

27:59you shift the

27:59dates to the last day is actually today

28:02because if you don't do that you'll see

28:04null for quite a lot

28:05of the Superstore data because it stops at

28:07a certain point in time depending on the

28:08Tableau

28:08release so let's go back that is that is

28:11pretty much the whole entire experience in

28:14a nutshell

28:14that is how to use the the the platform

28:17what I will say is this if I go back to the

28:19Tableau

28:20past landing page now that I'm here you'll

28:23notice that this summary here at the top is

28:26being

28:26generated so it kind of tries to pull out

28:28the most important things happening so it

28:30says this month

28:31there is a 5.9 percent decrease in

28:34subscribers um based on this video so what

28:36is Tableau explained

28:37in under 10 minutes basically it's saying

28:39hey there's a there's a six percent

28:40decrease in this

28:41one video compared to last month however

28:44there is a significant increase 28.3

28:46percent in overall

28:47subscribers compared to the same time last

28:50month so in broad brush terms I'm doing

28:52well but being

28:54specific there's one video that's got an

28:56impact so I might choose to go and analyze

28:58that I might

28:58choose to go and do something else now

29:01unlike other features in Tableau these

29:04thumbs up and

29:05thumbs down help improve the system Tableau

29:07is using these as prompts to try and

29:09understand what

29:10people are interested in and what they're

29:12not so if I say that this is this is good

29:14um you can

29:14obviously send that feedback to Tableau I

29:17won't bother here but um what this will

29:19help you do is

29:20try and understand what's going on the

29:22other thing to bear in mind you'll see this

29:25says it was last

29:26updated 12 hours ago and that's because I'd

29:28already created this data set whilst doing

29:31a demo so

29:31this this sentences this sentence suggests

29:34a couple of things and it's sort of a

29:36really intriguing

29:37fact about Tableau Pulse and how it's

29:40working behind the scenes 12 hours ago I'd

29:44just finished

29:45recording what was the third take of this

29:48video right and what's ended up happening

29:50is it somehow

29:51cached that data source and it's remembered

29:53that it's seen this in this data source

29:55before

29:55likely because it's linked to my Tableau

29:58cloud instance so rather than regenerating

30:00all the

30:00insights it actually found that it already

30:02generated them and is simply serving them

30:04up

30:04from the cache even though you've watched

30:06me create these metrics for the first time

30:08and

30:08add the data source for the first time to

30:11Tableau Pulse whilst doing this so it does

30:13suggest a

30:14little bit of caching a little bit of what

30:15I would suggest sort of frequency in which

30:17it's going to

30:18be updating these metrics based on how

30:20often the data source is updated so if the

30:22data source isn't

30:23updated I don't expect this to change but

30:25if it's updating frequently then I would

30:26expect Tableau

30:27Pulse to hopefully pick up on this

30:29frequency anyway these are small nuances to

30:32bear in mind

30:32as we start to get to know this feature a

30:35little bit more but nonetheless I think

30:37that is pretty

30:38much all I want to cover today we've been

30:39going for a long time 30 minutes it's a

30:41really long

30:41time for a video I really do thank you for

30:43watching this if you've watched it all the

30:45way through

30:46let me know what feedback you have about

30:49this video about the feature below I do

30:53know the

30:53Tableau product team do see the comments so

30:55go ahead let them know what you think save

30:58me the

30:58the sort of job of having to summarize it

31:00for them but the other thing to bear in

31:02mind here is there's

31:03so much more to discover so if you've

31:04discovered something you find it super

31:05useful if you've got

31:06some questions and you'd want to sort of

31:08find out you know how to how to how to

31:10execute that

31:11let me know in the comments below we'll try

31:13and see what we can do in the coming days I

31:14need to

31:15spend a bit more time generating a bigger

31:17range of metrics and a bigger range of

31:19insights so I'm

31:20slowly working through building a bunch of

31:22data that are going to slowly update over

31:24time and we're

31:25going to try and use Tableau Pulse to see

31:27how well it passes out some of that

31:28information that's super

31:30important in this particular case because I

31:32think this is a new feature it's going to

31:34have a lot a

31:35lot of sort of strange edges and rough

31:38edges that we need to sort of iron out the

31:41last thing I'll

31:42say though is that at least on this release

31:45now that it's public there are a few

31:47features missing

31:48from Tableau Cloud that I would have

31:49expected to be in here specifically if you

31:51're paying for

31:52something like a data management add-on I

31:53would have hoped to have seen more

31:55information about

31:56lineage if I go to a metric for example I

31:59would really hope that once I'm here I

32:02start to see a

32:03little bit of information relating to

32:05lineage maybe some charts and dashboards

32:08that use the

32:08related metrics so that I can go off and

32:11see the same context inside of the Tableau

32:13Cloud world now

32:14I do know that it's a deliberate choice by

32:16Tableau to make Tableau Pulse a separate

32:18experience so it

32:19doesn't have to be encumbered by all the

32:21effort and all the sort of legacy that is

32:23Tableau Cloud

32:24and Tableau Server but at the same time I

32:26think it's a normal expectation especially

32:28in the most

32:29recent years where Tableau and have been

32:31really pushing the data management add-on

32:33to lots and

32:34lots of customers to really bring that

32:35value through to new platforms otherwise it

32:38does sort

32:38of beg the question of you know you know

32:40when you when you subscribe to these things

32:42how secure is

32:43your investment if they're not being pulled

32:45through to new features you know if you'd

32:48used

32:48ask data or metrics in the past and you've

32:50spent the last three years finding a sweet

32:52spot where

32:52these work great I do think Pulse is a

32:55better product than metrics definitely but

32:57all the less

32:58you know the time you've invested in that

33:00kind of you know doesn't really pay off

33:02here it's a

33:03completely different experience you have to

33:05reset your expectations you have to rebuild

33:07you the way

33:08it works the other final thing is that we

33:11will do a video on how to use this as an

33:14end user some of

33:15the feedback I get is that look a lot of my

33:16videos are great for authors but what about

33:18just end users

33:19who don't care about building these things

33:20I'm just going to come and use these

33:22metrics we'll do

33:23a dedicated video on that just to make it

33:24easy for everyone to be able to share that

33:26as well okay

33:27we've been going for long enough and I

33:29really hope I haven't missed anything that

33:31's sort of my biggest

33:32nightmare this is the fourth time I've

33:34recorded this and so I think I've done it

33:36in a much more

33:37logical order this time around so this is

33:39just you know an honest honest experience

33:42let me know what

33:42you think let me know if there's anything

33:44missing whatsoever we can cover it later on

33:45this week

33:46as ever thanks for watching and I'll see

33:48you in the next video

33:49you

33:49you

33:59[ Silence ]