Tableau Data stories - New in Tableau 2022.2

0:00In Tableau 22.2 Tableau wants to tell you

0:03stories. There's a new feature that builds

0:05a story based on

0:06a data visualization inside of your

0:08dashboard and in this video I'm going to

0:09show you how it works

0:11and also explain a bit of the history about

0:13where this feature comes from. Let's get

0:15stuck in.

0:16Okay so we're here on Tableau Online I'm

0:18going to do this feature using WebEdit. You

0:20can do this in

0:21desktop but for this pre-release I can only

0:24really do this via WebEdit and I'm

0:26recording this before

0:27the actual full release just to get ahead

0:29of time. So I'm going to do this all

0:31entirely through the

0:32browser. The other thing to note is that

0:35for this feature to work the data source

0:37has to be a

0:38published data source. You can't just do it

0:40off a data source inside of the workbook it

0:42has to be

0:43a published data source. I'm not entirely

0:45sure what the restriction there is but in

0:48essence it

0:49you know it could be one of two things it

0:50could be there's a specific bit of

0:52architecture behind

0:53published data sources that allows Tableau

0:55to do this or it could just be Tableau

0:57putting a little

0:58bit of weight behind their own data sources

1:00to try and encourage people to do things in

1:02a much

1:02sort of more concise way. So I think it's

1:04more of an architectural thing than it is a

1:07Tableau trying

1:08to dictate how you do your work but

1:10nonetheless let's hop into one of these

1:12sheets and we'll first

1:13address the issue of a published data

1:15source. I think this workbook is actually

1:17got embedded data

1:18sources so let's go ahead and edit this and

1:20the easiest way to do this is just to go to

1:23any one

1:23of these sheets and if I go into the sheet

1:25you'll see that I have three data sources

1:28there and I can

1:29just right click on any one of these save

1:31as a published data source and because I'm

1:32doing this

1:33in web edit this should be almost

1:35instantaneous because I'm technically doing

1:37it on the server

1:37that's running the data source. So when I

1:40do that what I want to do is tick this

1:42option here

1:43to update this workbook to use the

1:45published data source so it will publish

1:47the data source

1:47then it will flip the connections around

1:50and so this connection over here will

1:52become a published

1:53data source. Let's go ahead and do that and

1:56let's hit publish hit okay and what we

1:59should immediately

2:00see is that the data source updates and as

2:02soon as I do that you can see that this

2:04icon over here

2:05on the left hand side changes to a

2:07published data source icon so that's worked

2:09and that's pretty

2:10much good to go. I can actually do this

2:13with everything in the workbook so let's

2:15just go ahead

2:16and very very very quickly do that just in

2:18case anything else comes up. Let's just do

2:20the same

2:20thing let's publish everything and when I'm

2:23doing this of course I'm editing this and

2:26there's

2:26actually a new feature that saves drafts as

2:30we're doing this so you can see that the

2:32the draft

2:33moniker is over here at the top and autos

2:35ave is also on that's a relatively new

2:38feature again in

2:39this particular release so there's two

2:41things that are happening at the same time

2:43in this demo. I've

2:44actually had an error on this particular

2:46one I bought in the replacement selected

2:47data source

2:48with a published data source so I think it

2:51did publish it but it didn't actually

2:54switch the data

2:55sources over that's annoying that's

2:57slightly frustrating so let's try it again

3:01and let's see

3:02if it works again and let's call this v2 so

3:04it doesn't publish over the previous one

3:06let's hit

3:08v2 publish success same error again it's

3:11not clear why this is you can copy these

3:14error messages and

3:16these aren't the most descriptive errors it

3:18might be something to do with the data

3:21source type or

3:22whatever so for now I'll just leave that

3:23one in there let's try it one more time

3:25with this top one

3:27see if this is actually a general bug as I

3:29said before I am doing this on pre-release

3:32software so

3:33I do expect bugs here and there this one's

3:35worked absolutely fine you can see that's

3:37published

3:37so we're just going to ignore anything that

3:39uses the sales target in this particular

3:41demo

3:42because we know that they're not using the

3:44published data source so now that I've done

3:46that can go back to the overview I'm still

3:47editing my workbook and I can actually

3:49start looking at

3:50this feature I'm going to find a scatter

3:52plot to start with because I think that's

3:54going to be an

3:55interesting one to start with because

3:56scatter plots are one of those things where

3:58they're a chart that

4:00when you look at you can generally see the

4:02trends but what I want to see is what the

4:04data story

4:05capability is going to add to the ability

4:07of reading the chart because for a lot of

4:10people

4:10the scatter plot you tend to read it at a

4:12very high level and I want to see if the

4:14data stories

4:15pulls out any sort of strange trends and

4:18outliers inside of this particular data set

4:22so in order to

4:23build a data story there's this new object

4:25over here on the bottom left called the

4:27data story

4:28and when you drag it in you can just grab

4:30it and drag it in you can see that I get

4:33this option to

4:33place anywhere on the dashboard like a

4:36dashboard item and when I drop it in what

4:38actually happens

4:39is a pop-up appears now I actually tried

4:42this earlier on and I have security

4:45settings on my

4:46browser and the browser actually blocked

4:48this pop-up from appearing so if you're

4:50trying to do

4:51this and nothing is happening just check

4:53that your pop-up blocker isn't blocking

4:55this if it does what

4:56you'll tend to see is something in this

4:59very top area here in your browser bar just

5:02saying look I've

5:03blocked a pop-up do you want to allow it

5:05click on that hit allow rerun the action

5:07again and it

5:07should open absolutely fine but it opens up

5:09a separate window the other thing is

5:11depending on

5:12your browser and depending how sort of

5:14savvy you are or aren't with your browser

5:16what could happen

5:17is a scenario like this where you're

5:19working full screen it's opened up and you

5:22think that it's not

5:22working because essentially what happens is

5:25the window is actually loaded behind the

5:27current

5:27active window so just double check behind

5:29your current window to make sure that this

5:32window is

5:32actually appearing it should load on top

5:34but again I've also had one instance where

5:36it didn't do that

5:37it disappeared and I went off did something

5:39else came back and completely forgot where

5:41it was and

5:41I thought it was bugged it was just behind

5:43the window okay so now that we've got the

5:45data story

5:45element here I'll bring it in right in

5:47front of the screen here so we can just

5:49focus on setting

5:50it up and when you go and drag that item in

5:52the first step is configuring it and this

5:54configuration

5:55step takes a bit of time and I'll come to

5:58that point a little later on but you

6:00essentially have

6:01to choose the chart that you want it to

6:03work on so for this one I want the scatter

6:05and once I've done

6:06that I hit next then it will ask me about

6:08the fields and for the fields I can choose

6:11the

6:11dimensions and all the measures now in this

6:13particular case I've got profit and sales

6:16so I

6:17might take the profit ratio out of this

6:19because the profit ratio is where the

6:21coloring is happening

6:22and the scatter plot is really just looking

6:24at profit and sell so I'm going to take

6:26profit

6:29ratio out of this and I'm going to move it

6:30to the ignored area just here at the bottom

6:32you can see

6:33this bottom area is the ignored area and so

6:35it's just going to look at customer names

6:37some are

6:37profit and some are sales let's go ahead

6:39and have a look at that so let's hit next

6:41and then the very

6:43last step is the story itself it's saying

6:46okay what story would you like me to build

6:48and for this one

6:49you can actually see it's loaded a story in

6:51the background but I can actually select a

6:54scatter

6:54plot in this particular case and what you

6:56want to do is use this particular one for

6:59relationships

6:59between numerical variables now notice as

7:02soon as I get to this tab this final story

7:05tab just right

7:06here and the visualization actually updates

7:09specifically the text here updates and so

7:11as I

7:12cycle through these selection it's giving

7:14me a flavor of what those text is going to

7:17look like

7:17it's actually starting to generate the

7:19story you can see it say writing story as I

7:21click between

7:22these different sections and so it's using

7:25whatever selection you've got here to

7:28change the way the

7:29story is done and you can see it's quite

7:31verbose we'll come to that again as well a

7:33little later

7:33on so it's really interesting capability we

7:35're just going to choose scatter plot

7:37because I think

7:38this one has the least text and I actually

7:40want that for this demo and we'll come to

7:42that again a

7:43little later on there's lots to unpack with

7:45this feature it's super interesting so let

7:47's hit done

7:48and that's it that's we've got our story

7:50now what I find a little bit distracting is

7:53that this data

7:55story is treated as its own object and I

7:58think it actually creates a bit of a burden

8:01of work on the

8:02author who has to put in a dashboard

8:05because it's not apparent to which

8:07visualization if I if I

8:09deselect the charts on here if I just go

8:12and select this little white space and I

8:14yeah keep

8:15my face here on the bottom right let's put

8:16it on the top right here if you just look

8:18at this

8:18dashboard it's really hard to ascertain

8:21what this text is relating to is it for the

8:24chart at the top

8:25the chart on the left the charm on the

8:26right I deliberately placed in the middle

8:28to create that

8:28confusion admittedly even if I'd place it

8:31to one side the text is going to be

8:33interacting with at

8:34least two charts and so it's hard to know

8:36exactly what it's looking at now this whole

8:38dashboard does

8:39look at profit and sales and profit ratio

8:42but again it's not obvious which analysis

8:44it's doing

8:45and where because and of course the

8:47narrative will be largely based on

8:49dimensions as well as measures

8:51in our data sources and in the scatter plot

8:54the data points relates to customers which

8:57is fine and

8:57on the right they also relate to customers

8:59so which one of these two charts is it

9:01talking about well

9:02the way you have to really solve that is

9:05the only way I sort of think design-wise to

9:07solve this is to

9:08put a container in there before I do

9:10anything with the container I'm just going

9:12to go ahead and put a

9:13border around it and this is not the prett

9:15iest solution but I'm just going to put a

9:16yellow

9:17border just so it's super clear and now

9:19that I've done that I'm going to put this

9:22inside of it

9:22and that inside of it okay what I would

9:26have liked potentially is this ability for

9:31the story to be

9:33this sort of additional information pop-up

9:36rather than be a separate object related to

9:39the chart

9:40and there's probably going to be a way to

9:41do that but this is why I said earlier on

9:43this creates

9:44quite a burden for the author because what

9:47the author now has to do is use the you

9:49know the tools

9:50that Tableau give you things like the

9:52ability to show and hide containers buttons

9:55and essentially

9:56use all those now features that used to be

9:59hacks to make this work nicely and it's it

10:02's it just

10:02makes the feature fall a little bit short

10:05in terms of usage what I would have loved

10:07is just for this

10:08to be an information icon you float it next

10:10to the title and when a user hovers over it

10:13the story

10:13comes up or clicks away and goes back

10:16rather than this what feels like archaic

10:18way of putting text

10:19on a dashboard and that's what it really is

10:21it's quite a lot of text and it just sort

10:23of goes up

10:24and down but let's re-change the balance of

10:27this dashboard here I think these are set

10:30to distribute

10:31evenly let's just double check that that is

10:33not the case let's do that and then we can

10:36move this

10:36around a little bit and give the text a bit

10:39more space okay so that's my sort of two

10:41cents on the

10:42initial sort of first impressions you want

10:44to put things in a container make them the

10:47same color so

10:47people know that those two things are

10:49related and it's super obvious now once you

10:51've put a story

10:52into the dashboard there's this little

10:55toolbar at the very very top and this is

10:58again twofold you

10:59can think of this as an author and you can

11:01think of this as a user we'll come back to

11:03what the user

11:03sees a little later on towards the end once

11:06I've created a story I can go into settings

11:08and actually I can actually enhance the

11:11story a little bit more now what I'm really

11:14getting here

11:15is fine-grained controls of how the story

11:18works I can make it less wordy or more word

11:21y I can allow

11:23users to change this particular setting I

11:26can actually go into these patific

11:28dimension terms

11:29and measure labels and change exactly how

11:32they're being used so you'll see that it

11:35says sum of profit

11:36across all of this I'm actually just going

11:39to use profit and again the burden of work

11:42here is on the

11:42author I'm having to remove things that are

11:44quite straightforward and I would hope it

11:46would do this

11:47a little bit better and what that what that

11:50leads to is essentially I think a slightly

11:52nicer way of

11:53reading this say as profit increased sales

11:55increased based on the data provided okay

11:58nice trend

11:59specifically when profit increased by one

12:01sales increased by 1.95 it's basically

12:04trying to give

12:04like a correlation to this there may be

12:07other factors contributing to sales but the

12:09year is

12:09evidence of a strong relationship so

12:12increased so basically what it's saying I

12:18just thought about

12:20that for one second I thought maybe it

12:21should have said this the other way around

12:23so for this

12:23particular section it should have said when

12:26sales are increased by 1.95 sales goes up

12:28by a factor of

12:29one now that doesn't make as much sense but

12:32basically saying it backwards makes no

12:35sense

12:35because profit is the end goal we want to

12:38make profit as a business so it's saying

12:41when profit

12:42goes up by one sales have to double

12:44basically to increase profit you know it's

12:47there's something

12:48about that sense that just doesn't seem

12:51right but that's because again maybe this

12:53is machine

12:54learning ai all that wonderful stuff and so

12:56it's going to take a little bit of time to

12:58get out of

12:58the uncanny valley that this stuff kind of

13:01tends to create now a few entities deviated

13:04from the

13:04general relationship indicating a good fit

13:07so what it's really talking here is that it

13:09's calling the

13:10customer's entity so you can go in here and

13:13replace this with customers um customer for

13:17one

13:17and the customers for uh that and you can

13:21add additional terms so uh you can say user

13:26and

13:26users okay and by adding these terms what

13:30you're essentially doing is adding sort of

13:34context to the

13:35glossary so as we look at this um a few

13:37customers deviated from this general

13:39relationship indicating

13:40a good fit the 793 users can be organized

13:43into two distinct groups of similar profit

13:46and sales values

13:47and so what it's doing is it's mixing up

13:49its use of these things i think it's

13:51literally alternating

13:52its usage throughout the sentences there

13:54were 760 customers that had values of

13:57profit so it's it's

13:58switching different terms so you might have

14:00lots of terms for the same thing you could

14:02use either

14:03and i don't like the way it switches them

14:05if that makes sense i'd rather just refer

14:07to them as

14:07customers all the way through and then that

14:09's more consistent but you can see here how

14:11this narrative

14:12section here i have a little bit of control

14:14as to how the text is formulated and so

14:17this is um this

14:17is nice now if we go to the next section

14:19the characteristics we can open up the

14:21characteristics

14:22and what we're really talking about here is

14:25formatting how it's presented to me now

14:27when

14:27we're talking about profit and sales these

14:29are numerical values and in this case i'm

14:31dealing in

14:32dollars so i can actually just tell tableau

14:33look this is actually a currency we're

14:35talking about

14:36and it'll format it in the dollar

14:38equivalent and it will deal with dynamic

14:40decimal places and

14:41it'll format negative values as per the

14:44dollars typically formatted you can assign

14:47meaning to

14:48large values so whether they're good

14:50neutral or bad and choose how to aggregate

14:52a profit so in

14:53this case it's a sum rather than average

14:56you might have some measures where it makes

14:58more sense to do

14:59the average let's say you're analyzing

15:01average price the total price isn't going

15:03to make sense

15:03and therefore you might want to look at the

15:05average price rather than the sum sum sum

15:07price

15:08and again with sales we can go ahead and

15:10select currency leave it as a dollar

15:11hit some that's absolutely fine and hit

15:14save and when you hit save it actually

15:17closes that window

15:18down for you and you can start to see we're

15:20starting to change how this works let's

15:22open up

15:22the settings again again if you don't see

15:24this sometimes it can disappear into the

15:26background

15:27so just make sure you've got it in the

15:29foreground if you go to the analytics tab

15:31we currently don't

15:32have analytics for this particular story

15:34type for scatter plots so we'll come back

15:36to that another

15:36time drivers and this is essentially how

15:41the the data source is perceived by the

15:45tool say source

15:46data must be either continuous or discrete

15:48source data must be either continuous or

15:50discrete so in

15:51essence this this particular section is not

15:53being driven by anything when we switch to

15:55a different

15:56chart type you'll see this may be working a

15:58bit better relationships and select

16:00measures that have

16:01relationships with other measures in your

16:03chart so actual versus benchmark current

16:05versus most

16:06versus period again these are just settings

16:08to really fine tune everything now the last

16:11thing

16:11is bullets and paragraphs so if you select

16:14bullets you get bullet points if you select

16:16paragraphs

16:17what is strange is it seems to me that it

16:19just collapses all the text again this

16:21could be a

16:22formatting thing we might get sort of

16:24spaced out paragraph more properly once the

16:27final release is

16:28done and that's sort of the general gist of

16:31these particular settings on the left hand

16:34side a lot of

16:35fine-grained control as to how the story is

16:38presented and now there's one more thing

16:40which

16:41is the edit tab this edit tab here allows

16:44you to edit this story a little bit more

16:46and this time

16:47it's specifically editing what's in front

16:49of you so the settings change how the story

16:52is built the

16:53editing changes the story itself so if we

16:56go ahead and select edit we get another

16:58pane and this time

17:00you can literally edit the story so the

17:02first options you have are the ability to

17:04add notes

17:05both at the top here and at the bottom so

17:07you can add additional text notes to the

17:10story and then

17:11the second thing you can do is you can add

17:14some general items so you can add custom

17:17items think

17:18of these as bits of things that you know

17:21you want to add so when i clicked on this

17:23add custom item

17:25it allows me to add a little sort of text

17:27range here and this is good in the event

17:30that maybe

17:30you've done your own stories i've actually

17:34got a customer where we add text notes to a

17:36report and

17:38this data stories feature might allow me to

17:40do that because it will allow maybe the

17:42first initial

17:43piece of information to come from the

17:45visualization and update in real time with

17:47the data source but

17:49then allow me to add text notes for more

17:51static things that maybe only change

17:53quarterly or monthly

17:54and so that that will work really really

17:56nicely just being able to add our own text

17:58first now you

17:59can make any of these visible or invisible

18:01you can see the little eye icon over here

18:03on the left hand

18:04side right hand side even and when you

18:06click on those you're just showing and

18:08hiding the specific

18:09story so now if i hit save you'll see that

18:12um if i uh take that one there you'll see

18:15that i basically

18:16only get one bullet point here um it's just

18:18regenerating it it's not changing the

18:20stories

18:21you're just regenerating it the the text

18:23and the information is not lost the

18:24analysis stays exactly

18:26where it is let's let's add a note here at

18:29the top um add a starting note and we'll

18:32add an ending note

18:34and an ending note uh hit save and you'll

18:38see those uh two notes that i've just added

18:42come up

18:43let's add a starting note and add an ending

18:45note i also can't uh type or spell um but

18:49nonetheless

18:49um that's it let's add a custom note uh

18:52what i could do is here start typing lorem

18:54ipsum dollar

18:56uh i'm stuck uh i run out of uh ideas there

19:01so let's just hit save i managed three

19:05words of

19:05lorem ipsum i'm sure there's people out

19:07there who can do whole paragraphs and there

19:09you go it comes

19:09up as a bullet point as if it was part of

19:11the narrative so this is super interesting

19:14um it's

19:14really nice to see this here and now let's

19:16try and add a different type of story on

19:18another tab

19:19to a different chart okay so let's hit save

19:22let's go ahead and close that and instead

19:25what we're

19:25going to do is we're going to build a

19:26dashboard so i want to use this forecasting

19:28chart that's

19:29included in superstore sales so let's go

19:31ahead create a new dashboard drag in the

19:34forecast

19:35and now that we've dragged in the forecast

19:37so i'm actually going to delete the legend

19:40over here on

19:40the right hand side drag the data story and

19:42we'll go through the process again this

19:44time a little

19:45bit faster so you can see that you can see

19:47that there's only one chart forecast and

19:49the fields in

19:50this particular chart are to do with the

19:53months and segment sales uh summer sales

19:55actually twice

19:56those sales and i guess there's three sales

19:58because what's technically happening is

20:01this

20:01chart has three segments so the sales

20:03appears three times okay that's why this is

20:06here it's

20:06basically telling me that um in in sort of

20:09a long-winded way so um we'll leave that as

20:12is

20:12because i mean technically it's interesting

20:16to see what this does because it's the same

20:20measure

20:20three times in different contexts so i

20:22think it will have to break things down

20:25based on the

20:25time perspective but also the segment as

20:28well so let's leave everything in there as

20:32we are and you

20:32can see here that it's suggesting that the

20:35continuous is the best fit here and um what

20:39is

20:39nice about this if i hit done you'll see

20:41that it actually breaks this down so you

20:43can see here at

20:44the top pivoted summer sales by period and

20:47then the breakdown goes into consumer then

20:49corporate

20:50then home office so it's actually taking

20:52these breakdowns and at the very very top

20:55we've got

20:55basically the high level summary and so

20:57here you can see there's a summary it's

20:59actually formatted

21:00in a slightly different way if we go back

21:02in to edit this you'll see that you get a

21:04little bit more

21:05uh sort of to work with here so if i just

21:07drag this out you'll see that i actually

21:10get now

21:10a section for each one i can actually

21:13switch off a section as well as enabling a

21:17section so

21:18all of this is fully controllable you can

21:21make this as busy or as light as you want

21:24it to be

21:25i'll come back to that in a second um but

21:27nonetheless everything is here that you

21:29could

21:29want to sort of hopefully modify or change

21:32um it starts to feel like a lot of work if

21:34you're going

21:34to use this to make a really sort of a

21:37logical thing and we'll come back to sort

21:39of the philosophy

21:41of this in a second but i just want to go

21:43back to the settings this time to see what

21:45new options we

21:46have if i go to the analytics pane you can

21:48see now we do actually have some capability

21:50in here that

21:51we can go through so you can choose

21:53analytics to write about in the narrative

21:55you can talk about

21:56correlation and you can obviously look at

21:59the breakdowns in the segment you could

22:01look at trend

22:02line analysis and you're basically setting

22:05these thresholds for different types of

22:07behaviors and

22:08when it should talk about them or when it

22:10shouldn't okay and this is again giving you

22:12fine grain

22:13control of you know how this works and the

22:16other thing you can do is if i go back to

22:18the drivers

22:19you can see that previously we didn't have

22:21any metric or dimension drivers now we do

22:24have some

22:24dimensional drivers and select the type of

22:27driver that has the greatest impact on your

22:29analysis

22:30whether it's the individual count the

22:32percentage change or the cumulative

22:33percentage if there's

22:35sort of a long view of something and again

22:37you can you can have a a click between

22:39these and you're

22:40basically changing these particular factors

22:43and so that to me feels like you really

22:45actually have

22:46to know your data very well as an author in

22:49order to do this and what what makes me

22:53what what is

22:54interesting about this is that i feel like

22:58out of the box this is a very hands-on

23:01feature this this

23:02this to me doesn't feel like a an easy to

23:05use out of the box experience because i

23:08think as an author

23:09i now have to spend more time first

23:11understanding the story it's telling me

23:14then figuring out what

23:15are whether i think my users are going to

23:17appreciate the story or not and then make

23:20changes to either get them closer to

23:22understanding the story or we're deciding

23:25that there's too many

23:26changes to make and then just wholesale

23:27removing it i think in reality that's

23:29actually what will

23:30happen people will have a go at this won't

23:32understand how it works we'll look at this

23:34video

23:34we'll look at other sort of demos we'll

23:36look at all these settings they have to go

23:38through

23:38and i'll go well this was supposed to make

23:40my life easier but it's now introduced

23:42six tabs that i now have to manage that

23:44just give a summary of my data now when it

23:48comes to the

23:49stories themselves i actually also think

23:52these are quite difficult to fine-tune and

23:54understand

23:55and it's just quite a text-heavy experience

23:58i was really hoping for something that

24:00would

24:01um be quite high level and aggregated for

24:04example cells were up by two percent show

24:07me the chart to

24:07show that and the you know values have gone

24:11from this to this which is 67 from this to

24:14this i can

24:15actually see this text over here and it's

24:17weird because you know tabla's always been

24:19a visualization

24:20tool but it just seems like in this feature

24:22we've sort of taken a step back and we've

24:24given people

24:24text and it takes up so much valuable real

24:27estate and it's a dashboard item which

24:30means people have

24:32to people have to decide whether or not to

24:34include it in their work whereas if it's

24:36something like

24:38explain data or ask data where it's part of

24:40a sidebar that makes more sense to me

24:42because it's

24:43then part of the chrome and users can take

24:45it or leave it as an as and when they wish

24:48and so

24:48i think i think this would have been a

24:50feature that was better as part of the

24:52chrome rather than

24:52as part of the dashboard by chrome i mean

24:55the interface tableau itself the sidebars

24:57the panes

24:58the pop-ups the things that tableau do for

25:01you as a vendor to make it easier for the

25:04user to

25:04understand what you're looking at rather

25:06than something that you as an author have

25:08to embed

25:08into your workbook then figure out how to

25:11space it then format it all of that work

25:13just takes so

25:14much time okay and um yeah there's just

25:16there's just a lot to really understand

25:18here so i'm sure

25:19as we use this as it gets enhanced over

25:21time we'll start to get a better

25:23understanding of how exactly

25:25it works but my general feeling on this it

25:29's it's like 50 50 this is nice but again i

25:33'm going to be

25:34intrigued to see whether this really gets

25:36used or whether it's going to still have

25:39low usage

25:39and whether this might then end up being

25:41pivoted by tableau to do something else to

25:44explain

25:44dashboards now there is a flip side to this

25:47and it's quite a positive one actually and

25:50that is

25:50when it comes to accessibility because what

25:53this tool is doing is essentially creating

25:56what i'd

25:57consider to be the closed captioned version

25:59of the dashboard it's literally taking the

26:02sheet and

26:02explaining what's going on in the chart and

26:05for people who are visually impaired this

26:08actually

26:08might be a great assistive tool to start to

26:10make it easy to understand what's going on

26:12in a

26:13visualization imagine an experience where

26:15um this technology was being used to

26:17describe what was

26:18going on in a dashboard or in a chart that

26:21would be so useful just based on click

26:23interactions or

26:24any other sort of selection interactions

26:26that someone would use it would then sort

26:28of enable

26:29this feature run this feature and then get

26:31it to tell a story of what's happening with

26:33the data

26:33in order to do that i think it would have

26:35to be a lot less sort of a boast than this

26:37but if it could sort of improve the way

26:40sort of highlights importance of things

26:43then i think that

26:43would work really well for example the top

26:45segment breakdowns here this is done really

26:47nicely the

26:48three series did not all move in a similar

26:50direction from december 2019 to december

26:532019

26:54november 2019 with summer sales two falling

26:58uh the most uh 5.09 percent and summer

27:01cells one

27:01having the same starting and ending values

27:04and some of those two finished uh trending

27:06upward in

27:07the final period what it's really talking

27:08about is the three different sales values

27:10for consumer

27:11corporate and home office so data stories

27:15it's a nice feature it's um good to have

27:18this was one of

27:19the key note sort of spectacular features

27:21this this i think is the headline feature

27:24of 2022.2

27:25but again i think a part of me wonders

27:29whether um this is a good example of table

27:34au just losing its

27:35heritage a little bit and you know as an

27:40author it disappoints me that this gets you

27:43know

27:44prominence over um something as simple as

27:48um you know uh conditional formatting

27:52inside of

27:52calculations we still have to do nasty

27:54hacks for that um it frustrates me that

27:56this gets sort of

27:57presidents over um improved um chart types

28:01for certain things for example up and down

28:04indicators

28:05just native as part of the product rather

28:07than us having to go around find shapes

28:10right calculations

28:10to do all of this you know the kind of

28:12stuff that makes it easy for me an author

28:14as an author to do

28:15this just isn't easy and for me to do this

28:18same thing in a better way using visuals

28:21just doesn't

28:22exist and instead we've got this sort of

28:24text-based feature so maybe i'm being too

28:27harsh maybe maybe

28:29i'm just um you know not the actual sort of

28:31model user for this feature but i i don't

28:35know it just

28:36doesn't land with me very well it's an

28:38impressive feature don't get me wrong but

28:41um i think it's got

28:42a bit of work and improvement to really

28:43find its place okay so it's now time to

28:45take a look and see

28:46what it looks like to the end user what i

28:48've done is obviously created two story the

28:50first one i'm

28:51going to go to the last one we looked at

28:53which is this one on dashboard number seven

28:55where we added

28:55the story to the right hand side of the

28:57forecast chart when you go through to it it

28:59obviously loads

29:00the story and this is what you get this is

29:03obviously a very text-heavy story so you

29:05get this

29:06very basically a simple text box you can

29:08just highlight everything now what i like

29:11about this

29:11is that there is this sort of um interface

29:14click um it's sort of not obvious it's this

29:16sort of blue

29:17highlighter that wasn't as obvious when we

29:19were looking at the dashboard item but

29:21actually it's

29:21a link and when you click on it goes into a

29:23deeper version of the story with a bit of a

29:25navigation so

29:26you can sort of click in and click out and

29:29again when you go down here you can see the

29:31consumer has

29:31its own click in capability and you can go

29:34in and get a little bit more depth when you

29:36click into

29:37those so there's even navigation into this

29:39and it just makes it like i said before

29:42quite a text

29:42heavy setup now there is a setting that

29:44allows you to control how much text shows

29:47up and lets the

29:48user choose how much text shows up if we go

29:50back to the general dashboards back when i

29:53created the

29:54scatterplot i actually enabled an option

29:56around verbosity which is now going to show

29:59up when i

29:59loaded this story so let's wait for this

30:01story to load and you can see here that i

30:02have this control

30:03here that gives me low medium and high verb

30:05osity if i just go to the other end of the

30:07extreme select

30:08high you can see that the text is a little

30:11bit longer now again i think this is a sign

30:14of this

30:14is a sign of lacking confidence in any

30:16typical um you know piece of text you'd

30:19normally say to

30:20someone that you don't need more words you

30:22just need better words right to get a point

30:24across

30:24that's why twitter is such a wonderful

30:26platform because it forces you in this

30:28constraint not to

30:29have more words but to have better words in

30:31the same amount of space and i feel like

30:33this verbosity

30:34thing is a little bit of a get out of jail

30:37free card um part of the feature set that

30:40allows the

30:40user to sort of dial up and down how much

30:44more text they want to have and kind of to

30:46me there's

30:47surely there's just an a right amount of

30:49text for the thing you're trying to explain

30:51um and and there

30:52shouldn't be really there shouldn't be any

30:54need to choose how many more words you want

30:56to see because

30:57it should just tell you what you need to

30:58know out the gate without any need to dial

31:01that up or down

31:01if it's useful it's useful and if it's not

31:03it shouldn't be there so that verbosity

31:05thing is a

31:06bit of another weird um thing to sort of

31:08bear in mind now the other thing i i sort

31:10of think is

31:11actually quite important is this copy

31:14button now this copy button sort of gave me

31:16mixed feelings

31:17the first one as an author something

31:19crossed my mind and just thought well wait

31:22how does an end

31:23user know that i as the author didn't come

31:25up with this there's nothing here that says

31:27this text has

31:27been generated by tableau data stories

31:29something to let you know that this is

31:31machine talking and

31:32not a human being who's actually crafted

31:34this analysis i think that's actually an

31:37important

31:37distinction because if you copy this you

31:39might then go and paste that information

31:41somewhere else

31:42and the fact was that copy button even

31:44exists that's been such a ideological

31:46sticking point in

31:47the tableau world which is letting people

31:49export your data and for a long time we've

31:52always sort

31:52of fought this battle of telling people no

31:55no don't export your data to this platform

31:58just go

31:58ahead and create a better experience of

32:00people stay in the tableau ecosystem and

32:02that's been

32:03the sort of champion approach to this way

32:05of thinking but now suddenly there's a

32:08massive

32:08copy button that just gives you this

32:10information for free and with no heritage

32:13as to where it goes

32:14i mean i could go and paste this in a word

32:16document and off i go and there's no sort

32:19of

32:19context i mean if i just open up um i just

32:22open up a notepad let me just open up a not

32:24epad here

32:25let's just paste it let's open this up and

32:28open this notepad here and we'll just drop

32:31this in

32:32we'll call this text and just see the raw

32:34text that it pastes so i can't even type it

32:38's so late

32:38so if i um put a new word let's do that

32:41this is what it pastes doesn't paste

32:45anything um that

32:46says oh this has come from tableaus you

32:49know machine learning wizardry it's just it

32:52's just

32:53text and so now this this text can go into

32:56any word document and the key question i

32:58ask you if

32:59you paste this in a word document or a

33:01powerpoint now how do you communicate to

33:04someone where this

33:04came from it's just a little bit difficult

33:07maybe maybe again i'm being a little bit

33:09backwards here

33:09maybe i'm just overthinking this and um i'm

33:12supposed to sort of approach it with a

33:14little

33:14bit more of a liberal approach um but i

33:16really struggle because the tableau this is

33:19the same

33:19company that is you know given me this sort

33:22of permission model um in in the same

33:25product by the

33:26way to get to the same chart i have this

33:29incredibly sort of detailed permission

33:32model for uh content

33:34and on the other side i'm just getting copy

33:37and paste for free um so again it's it's a

33:40it's a

33:40little bit of an interesting one i think i

33:42'd be really interested to see how this

33:43evolves over

33:44time i'd actually be interested to see how

33:46things like permissions um sort of interact

33:48with this can

33:49you disable whether that shows or not

33:51through the permissions model or something

33:53like that it'd be

33:54just really interesting to sort of find out

33:56more but a bit about that but nonetheless

33:58that's how

33:58it looks like to the end user now i said

34:00earlier on in the video that i just talked

34:02to you a bit

34:03about where this feature comes from and

34:06this is another great example of a recent

34:09purchased

34:09tableau have pivoted into a feature so

34:12tableau back in 2021 in fact this post was

34:15by mark nelson

34:16at the time he wasn't president and ceo but

34:19um it was a you know feature that was

34:21bought

34:23by essentially acquiring narrative science

34:25the narrative science had built this

34:26capability both

34:27on desktop and on mobile essentially this

34:30uh let's see if we can open the screenshot

34:32of this a little

34:33bit bigger so you can see here that

34:35essentially it was connecting to data

34:36sources and it was actually

34:37building some of these nice visuals i'm

34:40actually talking about getting as a user

34:42these uh simple

34:43big ass numbers as they call them and uh

34:46nice indicators telling you the trend of

34:48the data

34:49just making that kind of stuff really easy

34:51nice summarized stories with up and down

34:54indicators

34:54in the text this feels like a much much

34:57more watered down version of that and none

35:00of this

35:00sort of greatness has made it all the way

35:02in here um i i never use this feature well

35:04enough but i

35:05always remember being quite impressed and

35:07at the time i gave the same critique that i

35:09gave today

35:10which was this is great but make it easy

35:12for me to build charts and i will help

35:14serve my users better

35:16and so this is where this feature comes

35:18from it's really interesting to see this

35:20now as a headline feature and we've seen

35:22this with our stator in the past we've seen

35:25this with a

35:25bunch of other tools actually over time

35:27either tableau have copied them wholesale

35:30taken the ideas

35:31on board and enhanced them and done them

35:32better or they've just gone out and

35:34acquired a company

35:35that's doing that thing already and then

35:37brought it into the fold and that's what

35:39let's see how

35:39many year turnaround of this so this was

35:422021 december of 2021 and here we are

35:45almost uh

35:47not quite a year later but six months later

35:49and we have the um sort of end products

35:52that's been

35:54rolled into the features so it's really

35:56interesting it's really interesting to see

35:59that kind of quick

36:01turnaround um but again um as i said before

36:04um this feels like a feature that's moving

36:07away from

36:08tableau's heritage to me tableau's heritage

36:11is about seeing and understanding data not

36:13uh

36:15you know creating text boxes um that tell

36:17you about your data if that makes sense so

36:19that's a pretty long one it's late over

36:21here as well so um i've been sort of r

36:23anting and

36:24talking for a really long time let me know

36:26what you think i'd really love to know your

36:28opinions

36:28about this feature am i being too harsh do

36:30you agree with me do you not agree with me

36:32do you

36:32think that could be a better job with this

36:35feature are you going to take the time to

36:37create space on

36:38your dashboard for this feature that's the

36:40true test are you going to give valuable

36:42pixel space

36:43to this feature and if so let me know why

36:45thanks for watching and i'll catch you in

36:48the next video

36:49you

36:51[BLANK_AUDIO]