We have a specific standard for Waterfall charts within my organisation, and I have been struggling finding a way to achieve in Tableau without using a very messy custom SQL data source.
We have our main fact table loaded with Plan and Actual data with a dimension to differentiate. Maybe not the best start but this was how the consultants set it up!
Our waterfall wants to have the plan data as the starting bar and then show the steps of the waterfall for each division (A to D) (stepping up or down depending on whether they are over or under budget).
The waterfall ends with the actual bar and the steps in between essentially show where the overall variance to plan has been generated from.
The issue is that I need to essentially generate an two extra rows in my data - one for plan summed and one for actual summed, along with the split by division (plan-actual).
The way I have generated this in custom SQL involves a number of different group by combinations that are unioned together - not very pretty (we are using SQL 2005 so cannot take advantage of the new group by functions which would have made it much cleaner).
The issue with the custom data source is that because my Plan total and Actual totals are pre-aggregated, if a user say wanted to remove one of the divisions from the chart, those Plan and Actual aggregates are not affected - obviously no good!
Would be interested really in understanding whether there is any way that a union type of query can be performed within tableau?
Did you get to build this report in Tableau? I am trying to build a similar visualization and have trouble to do the same.