Normally I would agree with the WRS graph builder that it would not make sense to create a bar chart with a measure on the X-axis and another measure on the Y-axis.
However, I've got a user that would like to see the volume of deposits (the account balance - a measure) spread across interest rates (another measure). "This can be done easily in MS Excel".
I've identified two possible solutions that would work in WRS:
1) I can create another variable in the Infomap that is a copy of the interest rate, and set this variable to be a category. Might cause confusion for users since there will be two variables named "Interest Rate".
2) I can request that the user define some ranges (1-2%, 3-4% etc). This will to some extent reduce precision, and depending on the ranges it might not work across different products (credit cards, mortgages, deposits, penalty interest rates etc)
Both of these have drawbacks, but the ideal solution will allow the user to do this without requiring further work in the Infomap when they discover additional variables.
Any ideas for a better approach that would create the desired graph?
Just out of curiosity, why didn't you choose to create a scatterplot which allows users to build it with two measures? What are the drawbacks to your two solutions? Which did you like the best?
We have been talking about related ideas in subsequent releases so I'm interested in your responses.
Did even think about a scatter plot, so will see how that satisfies my user.
With regards to the solutions that we have identified, (1) is very much a last resort. I have made it my mission in life to try and reduce confusion amongst our users, so struggling to see how to make this obvious "at runtime".
As for (2) I have started considering an approach that dynamically divides interest rates into deciles based on the actual rates used. That might do the trick.
If we were to implement (2) I guess the groups would be robust over time, interest rates don't vary that much. It's just that my little geek inside tells me not to "hard-code"...
Scatter plot was an excellent idea. Was struggling at first, since the aggregation led to one blob in the report, the sum of account balance and average of interest rate (as per the aggregation rules in the Infomap).
When I discovered the "Aggregate or Detail" option in the data section (would be nice to have this for individual report elements too) everything was much better, and the scatter plot gave us pretty much exactly what we wanted 🙂
(We are, as you might be able to derive, following the "trial and error" approach in devicing our system)
The deciles is best solved in the datamart, and DI Studio is not my friend today (well, to be honest, it never is) so I'm strugglig a bit with the approach. Would like to build the code directly in the Table Loader in the mapping step rather than a custom code step, but guess I'll get there in the end!
Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.
If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website.