1 Reply Latest reply on Jun 19, 2018 5:22 PM by swaroop.gantela

    Blending data with filters but not able to see all info I want

    Amy Jao

      I'm trying to analyze which keywords are working the best for one of our marketing programs.

      I have 2 different data sources (revenue & cost) and they're at the different granularity so we're trying to create a view that I can see how each keyword is performing per month in terms of the commission, cost, profit..etc.


      Since it's 2 different data sources, I chose the cost data as a primary dataset and blend the revenue data which is used as a secondary one. And since we wanna be able to check the stats by month, we also set up filters for "month" in both 2 different data sources. But here comes the question... if we filter out by month for the cost data first and then revenue, keywords not generating any revenue of the month would be filtered out and vice versa, for keywords not spending of the month would be filtered out despite of they're generating revenue if we filter out the revenue first.


      So I'm struggling to see if it's because we're blending data so can't really have all info correctly being shown in the view? Or there's a way around where I can just filter out the month of primary dataset and it could also pull out the matching info from our secondary dataset instead of having another layer of a filter to exclude the info I would like to see?

        • 1. Re: Blending data with filters but not able to see all info I want



          This may be too round-about, and I'm not sure if it will be feasible for

          your true data set, but maybe it can give an idea.


          I have two mock datasources: cost and revenue.

          (I didn't quite catch how they were different with respect to granularity.

          Please revise them as needed to match yours, and repost.)


          To each datasource, I added a Lookup sheet that contained all possible months.

          I then cross-joined to this lookup sheet on a calculated field of 1 using:

          CROSS JOIN with Tableau's join dialog


          So when I blended the sources, it was on word and the LookupDate,

          which they both had in common. And revenue showed up in February

          for Word1 even though there was no cost for that month.


          The revenue and cost shown were of the form:

          IF [Date]=[Lookup Date] THEN [Cost] END


          Please see the workbook attached in the Forum thread.