Microsoft Power BI has added over two dozen new custom visualizations in the last two weeks. One powerful new type of visualization is R visualizations, which allow users to leverage R code for predictive analytics without requiring programming knowledge. The document demonstrates examples of R visualizations in Power BI, including scatter plots of NFL stats created with a single line of R code, and correlations between measures that non-programmers can generate with drag-and-drop. These new visualizations enhance Power BI's business intelligence capabilities and make predictive analytics more accessible.
2. INTRODUCTION
July 23, 2017
We are entering the golden age of business intelligence. In just the last two weeks, a couple of dozen new
custom visualizations have been added to Microsoft Power BI. According to Gartner’s Magic Quadrant
report, Power BI has surpassedTableau in terms of completeness of vision (farthest dot to the right) and
these new visualizations show the power of Microsoft’s vision. I will show an example of the R visuals and
then list some of Power BI’s new custom visualizations at the end of this report. These developments will
yield important insights ranging from solid information to actionable strategic insights.
One new and very powerful class of visualization in PBI is R visualizations.These were originally designed for
programmers and required installation of R and R Studio. Microsoft has taken R a step further, opening up
predictive analytics for non-programmers.An analyst still has much greater flexibility using R programming,
but the direction BI software is heading is increasingly a self-service model. There is a game-changing
paradigm here as well: the ability to add interactivity. When coupled with slicers or cross-filtered with other
visualizations, this can be very powerful. Other very promising visualizations include auto-play that may very
well be a new phase in animation following the longtime standard gapminder animated bubble chart. The user
can press play and cycle through cross-filters or drill-down on attributes. Other promising developments are
visualizations that write narratives and voice-activated queries; though these visuals are in beta stage. It is an
exciting time to be on the leading-edge of these developments.
3. USING R CODE IN POWER BI
I will demonstrate a couple of R examples below and then combine them into one dashboard. First, we can use customizable R
capabilities already available to produce scatterplots (using NFL passing statistics from 2014 through 2016). After downloading R and R-
Studio, this took only one line of code. Any other R graphs can be implemented, and the bonus is no read functions are required once
the data frame has been modeled in Microsoft.
4. R-CODED VISUALIZATION
Now we can create a slicer on year and use it to see how relationships between these measures have changed
each season; the visual will update accordingly. Here is the same visual with all three years selected; very
powerful results.
5. R FOR NON-PROGRAMMERS
Here is the new correlation visual that allows non-programmers to reap the benefits of R. In this visual, the user
only needs to select the measures to be evaluated (or drag-and-drop them on the values pane).
6. R FOR NON-PROGRAMMERS (PT 2)
Next, we can build two more visualizations (this time selecting circle markers), one for the Top 5 QBs
and one for the Bottom 5 by using the Rank function (only slightly more difficult than drag and drop).
For example, this could show how top and bottom tier quarterbacks are performing and analyzing
the differences.
7. HERE IS THE FINAL DASHBOARD: EVERY
VISUALIZATION CHANGES WITH THE YEAR SELECTED
8. HERE ARE SOME CUSTOM
VISUALIZATIONS YOU MIGHT WANT TO
CHECK OUT