Small Multiples in Power BI
Introduction
A lot of Excel and Power BI users and bloggers have been
demonstrating the use of Panel Charts for many years, including me! I
like panel charts in that they are really useful, their layout is attractively
simple and yet they are based on a lie!
A panel chart is a column chart or a line chart, for example
showing two or more categories or variables but with one major difference: the
two or more elements of it are actually separate charts made to look as if they
are part of the same chart. Like this:
This is a panel chart because it shows the results of four
different companies at the same time but notice there is only one vertical axis
per chart. Hence the illusion is that this really is one chart. Actually, there
is a bigger version of this panel chart, comprising eight companies and a small
change to the vertical axis of one of the elements, Eni SpA:
In reality, showing yet another panel chart, the inputs
behind panel charts can be long and complex: the one that follows is relatively
simple but the inputs behind the first example is much more complex and even
involves using the Camera Tool (not shown here)!
Small Multiple
The previous examples all come from an Excel file, using
Excel techniques. Here is an example, then of a small multiple in Power
BI.
Firstly, the data: Prof Aswath Damodaran has released
his annual data updates and I have downloaded the data I need from him. For
example, I took his files for margin data for the USA, China, Global … and
combined them all using Power Query in Excel. Since I want to analyse
the data using Pivot Tables, Pivot Charts and so on, I unpivoted
the data set once I had combined all Regions/Countries. A tiny snapshot
of my unpivoted table shows this:
You can see the data set comprises more than 10,000 rows in its entirety.
Without further adieu, here is my small multiple of that unpivoted table:
Enabling
Small Multiples in Power BI
Small
Multiples are currently in Preview, as at the December 2020 update to we need,
first, to enable them:
File … Options and Setting … Options …
Creating Small Multiples
To create my small multiples, I did this:
·
I got my data, first of all
·
Create a Stacked Column Chart
o
Axis = Industry Name
o
Values = Value
o
Legend = Heading
o Small Multiple = Region
You can see my selections in the screenshot above which
shows my first small multiple.
Enhancing Small Multiples
That was so simple and it did what I was hoping it would do …
almost. Notice the scroll bars, both vertical and horizontal … the visualisation
is too large for my page. Even when I hide the Filter, Visualisation and Fields
panels, the scroll bars are still there:
Format … Grid Layout
Select your small multiple visualisation and then, using the
Format … Grid Layout dialogue, increase the number of rows from 2 to 3 and, in
this case, the vertical scroll bar is no longer needed:
In this case, there are just six Regions in my data set so
now we can see all of them in my 3 * 2 small multiple. Unfortunately, there are
many more Industry Names and even with a grid layout of 3 * 1, the horizontal scroll
bar is still there:
And in this case, the vertical scroll bar returns (although
I have hidden it from view here) because there are six charts in the
visualisation and just 3 on view at any one time … unless we make it a 6 * 1
small multiple:
It is getting difficult to read now.
Well, feel free to filter out, say half of the Industry
Names for one 3 * 2 small multiple and the other half for another 3 * 2 small
multiple …
Alternative Visualisations
Do we have to use a Stacked Column Chart for small multiples?
No! Here is a line chart version of my small multiple:
Just click on the Line Chart icon to change your
visualisation immediately. Do notice, however, it might not be so easy to read
in line chart mode.
I also tried a Ribbon Chart and a Stacked Area
Chart to see what they might look like and I suggest you do the same and
draw your own conclusions!
Summary
In summary, I like small multiples and even with my
relatively large data set, it copes well and can illustrate my discussion
points easily. As with most Power BI Visualisations, it is really easy to
change form one visualisation to another. It is really easy to change the
formatting of the visualisation, too. Overall, I will be using and recommending
small multiples for future visualisation and data analysis.
Download: Because I am using someone else’s data set
and importing it into Power BI using a query, there is no file to download for
this page. However, here is the link to Prof Damodaran’s January blog post in which
he shows you where his data are.
Duncan Williamson
11th January 2021
No comments:
Post a Comment