3 Replies Latest reply on Jun 27, 2018 1:40 AM by Ciaran Bource

    Side by side subsets

    Eric Viglotti

      Hello,

       

      I have a feeling this has to be doable with groups, sets or unions in the data source somehow, but I want to create a dynamic, user defined set of subsets that I can show side by side. Basically instead of having a dimension that has values of Furniture, Office Supplies, etc, I want a dimension we can just drag in that is a dynamic subset of values, but most importantly they would NOT be mutually exclusive. For example, a bar chart that has the following rows:

       

      Office Supplies

      West Region Items

      Bob Smith Customer items

       

      That way we can compare performance of different cohorts of items. I know I can do this outside Tableau by just taking my data, making a union of it 3 times and for each of the 3 items going into the filter, I can give it a static category name like "West Region Items" and apply a filter for just the west region and nothing else, etc. That's easy enough to do with Prep. But I would like to do this dynamically so I have a set of parameters for each of the 3 different dynamic subsets.

       

      Is this possible somehow?

       

      Thanks!

        • 1. Re: Side by side subsets
          Ciaran Bource

          Hi Eric,

           

          Would you have a workbook or a screenshot to illustrate what you want to do ?

           

          Just to be sure I get it correctly.

           

          Ciaran

          • 2. Re: Side by side subsets
            Eric Viglotti

            No problem, good point. Please see the attached. The idea is that I know I can easily make a filtered list for the sales and profit for the Consumer segment. I can also make a filtered list for the West region. The idea would be to make a list that shows each as individual items, like I am simulating in this dashboard page. The point is that this is KNOWINGLY double counting, meaning there are records that are in BOTH filters. This would be similar to other types of reports that show values for different time periods as separate rows such as YTD, Last 12 Months, etc where there is naturally overlapping, but the point is to compare different cohorts of transactions side by side as if they were just one dimension.

             

            And again, the idea is to make this as dynamic as reasonably possible. I know I could easily take this data, run it through prep and aggregate/filter accordingly and union them into one data source that would then look to Tableau to be exactly as I have. But the dynamic-semi dynamic part is really the challenge...

             

            Thanks!

            • 3. Re: Side by side subsets
              Ciaran Bource

              Hi Eric,

               

              I've been through your Workbook and didn't find any real efficient way to do this or to achieve what you want.

               

              Indeed the best would be to do some data preparation with Prep or even custom SQL.

               

              Also found these posts that are a bit old but relevant but they don't give us the perfect answer unfortunately.

               

              How to add two dimensions underneath each other

              Putting multiple dimensions on a table row?

               

              Ciaran