I was asked to create a chart which combines area, line and columns, which does not exist within the Power BI visuals currently offered.
I ended up finding a solution through R.
However, Power BI Services only allows Python/R scripts to run through a personal gateway. This is a limitation, as the visual is meant to be live in a Workspace App used across a larger organization.
So now I am considering setting this gateway up in an Azure virtual machine… Which though it would work probably will go idle unless I set it up to be a 24/7 VDI…
Problem with that solution
And the problem with that is it would rack up the cloud resource usage bill from Microsoft.
- I understand Python/R code for data transformations requires the personal gateway to be live, but does this matter for visualizations also?
- Has anyone succeeded in setting up a personal gateway through a VM for the purposes of using Python/R script without the need for manual intervention?
- Would it be possible to use one single purpose-configured VM for all dashboards-with-R/Python across Workspaces within an organization? (Python/R enabled VM-as-a-service, if you will.)
If all else fails, and I do mean if there is no other way, I am considering using the HTML viewer extension. I could probably fit an R Shiny visual into a HTML page and then embed that into Power BI. But that’s going to be quite the hurdle just because my pride doesn’t let me hack Power BI by overlaying two charts.
Please note I am forced to ask a theoretical question here, because the solution will require me to jump through a bunch of bureaucratic hoops before anything gets set up to even test this out. I will likely even have to lobby to sell the organization on the concept before that, as personal gateways by default are not allowed.