• MarkDataGuy

Power BI - App Source Visuals - Review 1 - Box and Whisker

Hi There,


Welcome to my first of what I hope will become many reviews of visuals that are available in App Source. As we know, PowerBI has lots of visuals, at the moment, there are 34 visuals available by default. Plenty of people will just select from those 34 visuals and produce amazing results. However, once you look into App Source, there are loads of additional visuals available. Some are paid, some are free, some are provided by Microsoft, some are 3rd Party. If you have never seen them, then start up PowerBI and select "More Visuals -> From App Source" from the Insert ribbon. If you are using the web service, then edit a report and select "Get More Visuals" (it is the 3 dots ... on the visualization panel)


Today, I am going to focus on the Box and Whisker chart, also known as a Box Plot. The Box Plot is a way of displaying information about the underlying distribution of data that you are analyzing. I have a worked example here. If you search App Source for Box and Whisker, there are two options available. "Box and Whisker Chart by MAQ Software" and "Box and Whisker chart" from Microsoft.


In my continued analysis of my home internet speed, I have tested both of these visuals against the results of my automated speed test, which captures speed results every 15 minutes. This is what I have found.


Box and Whisker chart by MAQ Software

Firstly, this is a 3rd party add-in. As with any third-party component, you should probably consider the longevity of the company, before you use their visualization in a report. I have no special knowledge of MAQ software. They look great, but please do your own analysis to make sure you are happy with them. PowerBI visuals have been known to disappear.


The MAQ visual is simple to work with and features some nice options to give you some good insight into your data. Under the fields panel, you have the following options.

  • Axis

  • Axis Category 1

  • Axis Category 2

  • Value

  • Dots Size

  • Legend

In the basic usage, you will need to use Axis and Value, and this is enough to create a basic chart, showing you the key metrics of a Box Plot which are;

  • Median

  • Upper Wisker Value

  • Lower Whisker Value

  • Quartile 1

  • Quartile 3

Plus the option of Average (Mean) and IRQ (the difference between Quartile 1 and 3). The following visual is what you get. You get immediate insight into just how poor my internet speed is, along with the speed distribution over time. The fact that the median (the line between the light and dark grey region) is lower than the mean (the white circle), tells you that my data is skewed to the left and my speeds are even worse than what the average suggests.


This is great, but when you start using the other fields, you can get much more insight. In my next example, I have used the Axis Category to allow me to view the speed over two different days, dots size to see if the distance of the test server makes any difference and the legend to identify the name of the test server.

Now when I compare the 2 dates. I can see that my performance on the 29th of Jan, was very different from the 28th, and by looking at the position of the Median line (between dark and light grey), I can quickly determine that it was actually much worst. If I was to judge this just by the average, you might think the performance was similar. Also, I notice that the distance to the test server doesn't make a difference, as I can smaller and bigger dots scattered across the distribution.


Finally, if I utilize the 2nd Axis Category, you can get an additional depth of information as follows;

In my case, this visual doesn't add any significant benefit as I just don't have enough data categories to analyze, however on different data sets, this could be very useful.


Box and Whisker by Microsoft

Onto the Microsoft offering. Initial impressions of this are that it is simpler than MAQ, but I found the MAQ one to be more intuitive to use. You have to use the three available fields to generate a basic box plot which are;

  • Category

  • Sampling

  • Values

The basic plot looks as follows;


If I was to compare this to the MAQ basic plot, the MAQ contains more information. Yes, I get the classic 5 box plot measures of

  • Median

  • Upper Wisker Value

  • Lower Whisker Value

  • Quartile 1

  • Quartile 3

and I can see the average via the dot in the middle, but I don't see the actual distribution of data in my plot.


There isn't much more to the MS box plot. I can select multiple entries of the category field to get multiple box plots as follows;

So now, I can compare my performance over different categories, however, I can't replicate my two category analysis that I was able to do with MAQ.


Conclusions

I like Box and Whiskers, and I would love to see more of them in use. Both charts discussed here are good and do what a box and whisker are supposed to do, however, in terms of features, the MAQ chart is better and I find that when I have had occasion to use one, I have used MAQ.


If you want to view my box plots in action, you can access them here. The PBI files and source are available on my git hub here.


Thanks for reading. I hope you found it useful.


Mark


33 views0 comments
Original on Transparent.png