Re-Upload - Metrics in Tableau Next Part 4 of 6 [Reposted]

0:00Okay, we're back.

0:01Today we're gonna build metrics.

0:03Now I should have strictly done this before building visualizations because I believe metrics can be used when building visualizations.

0:11And it allows you to define something at the semantic level.

0:15And that's essentially the step I missed at.

0:17So I went straight to building visualizations.

0:19rather than defining metrics that we're going to use in our visualizations and later on as well.

0:24So let me head to the Tableau Tim workspace.

0:27And once we're here, you'll see the visualizations are built yesterday.

0:30Here we need to go back to the superstore data model.

0:34And in here, there's an option there, select new, select metric, and we have it.

0:38Now there's a little bit of a rantab to get out of the way.

0:42Metrics used to be a feature available in Tableau.

0:45They were deprecated when Tableau Pulse was launched because Tableau Pulse was supposed to supersede Tableau metrics.

0:52Tableau Pulse is a very good feature.

0:55Has just about reached this sort of level of general awareness inside of the Tableau community, but then Salesforce.

1:00Salesforce in its wisdom decides to take all the best things out of Tableau Pulse, a thing you know and understand, and put it into a new capability in Tableau Next called drumroll metrics.

1:11So we're back to square one.

1:14So this is super confusing, super confusing.

1:17So I passionately hate the fact that these are called metrics when there's an identical looking product that functions better called Tableau Pulse

1:26Just why?

1:26Just just call it Tableau Pulse metrics everywhere.

1:29Full stop.

1:31Anyway, round over.

1:33Breathe.

1:33We're gonna create a metric.

1:34We're gonna call it profit.

1:38We're gonna call it profit, and what we are going to do is also put the description and call this profit as well.

1:45Hit next.

1:46Once we've defined that, we can go and get our measure.

1:48It's very similar to Table A prep.

1:49We can actually filter this here.

1:51So if you wanted to say hey, I want this profit to only be linked to this one metric or one dimension, you could

1:57Go ahead and filter it.

1:59You could also remove maybe something that's not meant to be part of that measure.

2:03That's quite useful.

2:04Let's say that you want to measure return on investment, but return on investment can't include

2:09um uh transactions from a certain dimensional attribute this would be a good way to sort of solve that so we're not going to do that here today I will say one thing this is this is kind of broken so when you select say city as a filter here

2:21It doesn't show you the list of possible subjects.

2:24So you could type something in here like elephant that might never match a piece of your data, but there's no sort of like failsafe.

2:32That doesn't seem right.

2:33You should only be able to select things that I think sit in the card.

2:36column.

2:36So um that's a little bit of product feedback.

2:39But you have operators you can use and you can sort of exclude things as well.

2:43So it's actually quite good that you've got that capability, but it's just

2:46I just wish you could see your data whilst building these metrics, otherwise you have to have a notepad somewhere and have taken it down.

2:52But anyway, um you can do more advanced setups, so you can set up sentiment aggregation, the type of aggregation.

2:59But we don't need to do any of that.

3:00This is just going to be a simple metric.

3:02The time dimension will be order date.

3:04And now that we've done that, we get a number.

3:06Now it's not doing any comparisons just yet because I don't believe we have set up any comparisons.

3:11So let's go to the next step where we can add dimensions that the user can filter again.

3:16So we can pretty much go to town, give the user all the flexibility.

3:21to filter the dimensional things they probably might want to so category I'm just gonna do the most logical ones uh subcategories we're gonna last one state and province that's that looks pretty good to me so then hit next

3:34Now that we've done that, we've got all of those dimensions added listed.

3:38You see that this list is actually sort of lengthened.

3:42So record level details.

3:43So the record ID in this particular case

3:48Strictly speaking, oh gosh, what level of detail is order ID on?

3:53It's actually at the product level because it's about products in a basket.

3:57So that is actually, in my opinion, the right answer

4:00So the record level details are there, the feels.

4:04Oh, this is interesting.

4:05Your metric starts with these recommended insights.

4:07I don't actually understand this, but there's a nice little information icon I should click on.

4:11Choose the field to use as the display name for each record, such as the customer name.

4:21This is weird because now I feel like I should have chosen the uh the the product ID.

4:31Choose the field that contains the unique system identifier for each product.

4:34records such as custom ID.

4:36I think this is right.

4:37It would be nice if you could choose like a a couple of fields because it might be that it's order and product.

4:43So I'd like to generate an order and product key.

4:46Again, I could go create the calculated field at the point of ingestion and bring that in, but I don't have that records level data

4:54So enable this to drill down and take action on individual records.

4:58If disabled, these actions are not available.

5:00Let's switch it on.

5:02Trends and changes.

5:04Record record level outliers

5:07Um define records unusually high, yes.

5:11Uh current trend, yes, yes, yes, yes.

5:13Contribution breakdowns, yes, yes, yes, yes.

5:15These are all the different types of analysis it can do.

5:17Um, a singular unit is um

5:20Um well it's the same, so it's profit.

5:23So you're just telling it what to call it, and it should also be called profit there again.

5:27Uh just do a capital P.

5:30Hit next and then let's look at insights.

5:34Oh no, that's it.

5:35We didn't even give it a comparison.

5:37But there it is, there's our profit metric.

5:39So we've got time dimension.

5:41uh additional dimensions filtering and we've got a description.

5:45Now when we click on it we can go back and change any one of these.

5:48It's kind of remembered all the settings as it should.

5:51So that's pretty much it.

5:53So oh this is interesting

5:55So earlier on when I clicked on these dimensions, it went to this tab, but I didn't actually recognize that it'd gone to that tab.

6:01So that was my apologies.

6:02These are the insight dimensions that it's going to use.

6:05We've chosen that.

6:06Okay.

6:06The thing I can't see here is how do I compare it to the previous, like, you know, whatever.

6:13It doesn't really say.

6:15So um I'm not getting a line chart because there's no comparison data.

6:19It might just be there's no data yet, actually, real data, because I didn't do a smart job with this data set and pin it to today's date.

6:27I actually have a tableau

6:29prep flow that does this for me in my Tableau Cloud instance.

6:33So maybe I should export that for this upcoming year so that I've got a good sample store per store sample superstore example.

6:40So that's profit.

6:42It's done.

6:42We can go ahead and create more.

6:44I don't think I need to show you that.

6:46Another thing you can do here is create a parameter.

6:48It's weird that this is done at semantic level.

6:50I would do this at the visualization layer, but it's good that it's

6:53it's allowed to persist here at the semantic level because it then affects everything that uses that semantic model.

6:58So it makes a ton of sense.

6:59Uh logical view um is one of those um

7:03Sort of combined tables that we had.

7:04So joins and unions happen at this lost sort of logical view, and data objects are just more data sources.

7:09So we've already gone pretty much to this full gamma of items.

7:13But if I go to my tablet in workspace and we look at this list, we don't see metric here.

7:20So this this is what I was wondering.

7:22Why do we not see a metric in this list?

7:25You would expect it to be um a child of of one of these or some sort of component, but it's just sitting there in superstore data model um ready to be used.

7:35Okay, so

7:36Metrics are pretty simple.

7:39You can uh you know kind of push these uh a little bit harder by building sort of more nuanced um

7:47settings but it's it's very similar to Tableau Pulse.

7:51There's really not much more to say other than that.

7:53Um so one thing I want to check if I go back to my homepage

7:57And um you see the metric just doesn't seem to appear anywhere.

8:01It's like I never created it.

8:03And if I just refresh the page just in case it's something

8:06to do with cash and I'm seeing uh here we go it was something to do with cash I'm glad I did that so now I can see the metric there if I go to the tableau space

8:15Nice.

8:16Uh do we see a metric here?

8:17Do we see a metric here?

8:19No.

8:19Let's go ahead and refresh this.

8:21See if this is also suffering from caching issues.

8:25No, I cannot see that.

8:27Um let's see if I can search all that is um so cool let's just search profit.

8:35You see it has dashboard, visualization, semantic models, data objects, but no metrics, you can uh not really see them.

8:43Okay, so maybe this is just not available yet.

8:46Um but yeah, it it's there.

8:47Trust me.

8:49Trust me.

8:49bro.

8:50Okay, cool.

8:51So we've built a metric to see this metric.

8:53I think the easiest way I'm going to do this is go and build a dashboard so we can start with that and then we'll end the video uh having just basically build the metric so

9:02go to add new dashboard and then the first thing we're gonna do first thing we're gonna add to this page is the metric now

9:13This is the dashboarding view and it's probably worth just explaining a bit of context around this.

9:19But you know, on reflection, no, this should be a whole separate video because I think the dashboarding experience is super important.

9:25So tune in tomorrow to see that.

9:28I will just generally go and see if I can add a metric.

9:32So what I did there is added a page.

9:34You can see add multiple pages.

9:35They essentially become tabs in your report.

9:38Makes a lot of sense.

9:39I don't need to

9:40many of these um how do i add a metric this is one of those weird things where if i go to the content

9:49There's nothing here.

9:50So if I go to oh here we go.

9:52So you have these icons here at the top, and they specifically call out to things you've built.

9:57So if I click on that and then I drop it in,

10:01I so what I do is I create the space, then I go add the thing, let's choose the dot matrix, select it.

10:09And then that's what adds the asset on.

10:12So I built it somewhere else, that's what adds it on.

10:14So to add the metric, I should be able to click on this.

10:17It gives me an identifier.

10:18I put it down, I select add metric.

10:21It asked me which semantic model to get it from.

10:23Good choice.

10:24Select.

10:25There's my profit metric.

10:26Select.

10:27Boom.

10:30There's my metric, and I can style this up and set it up however I want.

10:34Um, I should really probably add something a little bit more

10:38uh descriptive that shows off this capability because this has got loaded nothing basically and the data stop has stopped in July 23rd.

10:45So we need to go and get something a little bit

10:48better but long story short there's my metric and you can see that it's it's actually got quite a lot of descriptive information about the metric over here on the right

10:55hand side so there's our first dashboard.

10:58Um this was very brief it's very straightforward it's basically Tableau Pulse but not as well polished as Tableau Pulse.

11:04Tomorrow join in and we will start going with the visualization experience.

11:10Okay, thanks for watching, see you tomorrow