Microsoft recently updated the R version for Power BI Service and installed numerous additional R packages. For comparison, I posted the then-current list of packages back in July 2017. If you use R in Power BI Service, definitely see what’s available in the screenshots at the end of this post. Hopefully the Power BI documentation will be updated soon (it’s public on GitHub now by the way, and anyone can fork, edit, and submit a pull request). In addition to more packages, the new R version is 3.4.1.
Thanks to Anselmo Gomez for notifying me of a change yesterday. He happened to test out an R HTML custom visual called R Pivot Table that I created for BlueGranite (AppSource submission now pending!). While it worked in Power BI Desktop by installing the htmlwidgets and rpivotTable packages, it would not work in Service until the rpivotTable package was available. At some point this month, Microsoft installed it!
Available R Packages
Remember, simply because a package is available in Service, it does not mean that it is a supported package, a list of which appears in the documentation. Some of these packages will work with Desktop’s R visual, some will only work if you build an R HTML custom visual, and some will not work well at all in the context of Power BI. For example, I’ll challenge someone to get shiny to work as an R HTML custom visual. Although the package is available, that does not mean you can always use it without an error (this ultimately looks like a dependent httpuv issue, but in any case, I’m hopeful someone can get this going).
I tend to focus more on data visualization and geospatial, so some notable additions for me include shiny, crosstalk, leaflet, rgl, UpSetR, qicharts, sf, ggridges. Most or all of the tidyverse packages are now available, and so are newer ones for text analytics like tidytext.
Here are some screenshots of the updated packages (sorry, no text). Save and enlarge if the package names appear too small in this blog post.
Please post when R Pivot Table is available – it sounds excellent!
I received the first round of feedback last night from the Office Store team. It looks like it is going to be a balancing act between performance and trying to save a user’s state. I’ll post whenever it’s published. In the meantime, check for non-production releases on GitHub: https://github.com/BlueGranite/RPivotTable-for-Power-BI. The one submitted to the store includes the ability to save your selections, but it also made it noticeably slower.