The calculated field was created off the parameter I created [Measure] and the parameter contain sting, so no I don't believe there are any aggregate calculations in the field.
Well I just changed the parameter from string to numeric in the attached. Moved the calc from dimensions to measures.
It still works. In fact, I also changed the calc field from discrete to continuous on the filter shelf, and that still works.
So I need to see what Scott is doing to figure out why his won't work.
Scott, can you hack up a sample workbook that demonstrates what you are seeing?
1 of 1 people found this helpful
Anyone interested in this technique will also be interested in Robert Rouse 's new blog here:
Pretty cool effect.
People are finally starting to believe that it's okay to take objects outside the dashboard
Great job! The video & audio explanations were fantastic, and most importantly: I get it!
There's no doubt about it: this is an important skill I'll be using in the future.
I just re-watched this for the first time in a year or so... still helpful and relevant today! Cheers Joe, Ville, etc
I just watched this...several times. I think it needs to be redone. The non-sync'd mouse and interruptions made it difficult to follow along for me being a newbie to this technique.
It is an okay instructional but could really be improved.
It's been quite a while since the recording but I'm happy to walk you through it anytime via Screenshare. I just rewatched it myself a few times yesterday to review the popping technique, and had to pause, etc to follow every bit -- ignore the cursor you see, as its not Joe's.
If you (or anyone else) want to do a screenshare sometime we can always record it again - this was spun up in a matter of minutes.
I consider it to be better than anything else I've been able to find explaining the technique -- following Joe's clicks eventually led to my ah-ha moment. I'm hopeful it'll get more documentation eventually.
Yeah, recording stuff really isn't my forte.
And in the time since I recorded that, I've worked up some newer techniques too. Doing it with a slicker-looking workbook would also enhance the tutorial.
I'll have to work on that.
I realize that this is somewhat of an old post, however, I did want to compliment on the dexterity and the in depth detail and quality shown in the video. Good Job! (Yea...green lagging mouse was weird, but I forgot about it in a few minutes and just followed what was done.)
However, my experience with the same process has been a failure, not because of the idea, but because of how Tableau actually handles the data. Let me share why. First, know that I publish to Tableau server.
I created 1 dashboard, consisting of 7 worksheets from 4 separate data sources. All 7 of these sheets are filtering extracted data sets from as little as 1.5 million records for a worksheet, to 5.5 million records for a worksheet. And, of course they filter across all data sources (Using v10.0). The process takes about 45 seconds when changing the single select filter. The interaction is great, but speed is obviously not tenable.
Independently each sheet takes about 6.5 seconds to load. So, we already know Tableau is crunching hard. What the single dashboard does in our case, when changing the parameter as described in the video, is it loads the data for all 7 worksheets, even though you only see 2. Then when you again change the parameter, it again reloads the data for all 7 worksheets. So, you get the 40-45 second wait time every time you change the filter. This was further confirmed by running the debug thingy Tableau Desktop has (I forget the name at the moment)
For small to medium datasets, the speed would obviously be much greater, which is why this video was such a good illustration.
For a Large Database such as I have in this case, the option is multiple Dashboards; so we end up using the Tabs. This works ok enough, but I would prefer not to display each of the "secondary dashboards", and only display them upon request.
So I am sharing a different experience where a great idea does/did have its limits. Of course, if there were some magical trick where the process would only load that which is displayed, I would love to know what it is.
Chris -- Yup. It's definitely a limitation, and I have had to break things apart into separate dashboards at times too.
And yes, even though only sheets 1 and 2 are going to end up getting displayed, Tableau still has to render sheets 3-7 to figure out that nothing is going to get displayed.
Just a design idea here ... When I have to break into separate dashboards, I create a bunch of simple icon sheets (using a very small data source!) so that instead of users selecting a parameter, they click on little icons. And those have filter actions that go to the respective dashboards. All of the dashboards have the same layout, and they all have the same series of icon sheets so that users can click from any dashboard to any other.
I have 7 sheets/charts and have created parameter that selects these 7 charts.
I've decided to with the pop-route. So I created 7 pop sheets for the 7 charts. Then I created a floating container container consist one of the pop sheet and the charts.
In summary, i have:
- Floating Container 1: Pop Sheet 1 and Chart 1
- Floating Container 2: Pop Sheet 2 and Chart 2
- And list continues
All of the containers have the same width and height and place on top of each other in the dashboard and using the parameter as my controller for the 7 containers.
Now, the problem arise when I made my container selection for instance, I select Option 2 which points to Floating Container 2. But it seems seems that the Floating Container 2 is hidden behind the other Floating Containers.
How do I rectify this matter?
In desktop, you don't rectify it. But when you publish it to server the HTML behavior DOES let you "drill through" the upper containers to the one that is getting displayed if it's not the topmost container in the floating order.
Nice trick, Joe. Thanks to you and @Matt Lutton for sharing. I've been a user for about five years now and finally have an opportunity to spend more time building better functionality for the team. It took me a minute to wrap my head around it, but I got it to work in our scenario.