Not sure....but find my approach as reference below and stored in attached workbook version 10.1
b. Delinquent num:
IF [Loan Status]="Charged Off"
or[Loan Status]="Does not meet the credit policy. Status:Fully Paid"
THEN [Number of Records] END
project3_working.twbx 8.6 MB
Thanks Norbert. I actually had an answer from Stackoverflow before Tableau approved posting my question, but your answer really helped me understand what is going on. Thanks!
Actually, I think there is something un-"Tableauic" about my whole approach. I used to create dummy variables all the time in research but it does not seem to be very common in Tableau. It makes me wonder if there isn't a completely different approach possible with Tableau that I am over looking?
2 of 2 people found this helpful
I've been trying to figure out succinct ways to describe this difference over the last couple of years, here's another attempt:
When we're working with Excel or Python/R/SAS/SPSS etc. the workflow is to start out with a data set, then do a transformation of some sort such as creating a derived variable, using a SUMIF() function to do an aggregation, etc. that creates a new worksheet/table/data set, do 0...N more transformations where each transformation creates a new worksheet/table/data set, then finally graph the results.
Tableau has an "all at once" model where we start out with the Tableau data source (which might be a combination of multiple tables from multiple databases/worksheets) and when the view updates by dragging and dropping a pill Tableau queries the sources, does computations at up to four different levels (record-level, aggregate, table calculation, and post-aggregate) and potentially many levels of detail once we start getting into top & conditional filters, computed Sets, context filters, Level of Detail expressions, etc. and (finally) draws the view. Tableau has an internal logic around this, key elements are understanding the levels of calculation, dimensions & measures, order of operations, and the levels of detail. This all works fantastically in the background as we are dragging and dropping pills right up to the point where the way we've been conceiving of building out a view doesn't line up with how Tableau does and we get stuck because the details of what Tableau is doing are obscured (or completely hidden) from our view & discovery.
Norbert's solution identified the dimension needed on the vizLOD (the bin of the loan amount) and used that, while still using the delinquent num as a measure.
Thanks, Jonathan. That really helps me put it into perspective. I have found that getting away from the "Excel or Python/R/SAS/SPSS" way of doing things has actually helped my intuition and understanding. I really appreciate the way that Tableau users are willing to dive into the intuition behind how things work in Tableau, too.
Once again thanks for "stepping in" and taking a "step back" and elaborating on "behind the scene's" concerning the Technic and way of working of Tableau. Appreciated!:)