1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Much of what we would consider aesthetics has already been discussed above: things like alignment,
color, font choices, clutter. But there are a few more best practices for report design worth discussing
and these deal with the overall appearance of the report.

Remember that the function of your report is to meet a business need; not to be pretty. But some level
of beauty is required, especially when it comes to first impressions. Nashville consultant Tony Bodoh
explains "Emotion fires a half-second before logic can kick in." Readers will first react at an emotional
level to your report page, before they take more time to dig deeper. If your page looks disorganized,
confusing, unprofessional…your reader may never discover the powerful story it tells.

TDI blogger and TechTarget industry analyst Wayne Eckerson has a great analogy. Designing a report is
like decorating a room. Over time you purchase a vase, a sofa, end table, a painting. Separately you like
all of these elements. But although each individual selection makes sense, collectively the objects clash or
compete for attention.

Concentrate on:

Creating a common theme or look for your report, and apply it to all pages of the report
Using standalone images and other graphics to support and not detract from the real story
And applying all the best practices we discussed up to this point in the article.

Principles of visual design

We’ve looked at the principles of report design; how to organize the report elements in a way that makes
the report easy to quickly grasp. Now we’ll look at design principles for visuals themselves. And, in the
next section, we’ll dig down into individual visuals and discuss best practices for some of the more
commonly-used types.

In this section, we’re going to leave our example report page alone for a while and look at other
examples. After we’ve gone through the principles of visual design, we’ll return to our example report
page and apply what we’ve learned (with step-by-step instructions).

Planning – choose the right visual

Just as it’s important to plan out your report before you start building, each visual also requires planning.
Ask yourself “what story am I trying to tell with this visual?” And then figure out which visual type will tell
the story best. You could show progress through a sales cycle as a bar chart but wouldn’t a waterfall or
funnel chart tell it better? For help with this, read the last section of this paper “Visual types and best
practices” which describes best practices for some of the more-common types. Don’t be surprised if the
first visual type you pick doesn’t end up being your best option. Try more than one visual type to see
which one makes the point best.

Understand the difference between categorical and quantitative data and know which visual types work
best with what type of data. Quantitative data is often referred to as measures and it’s typically numeric.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 18/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Categorical data is often referred to as dimensions and can be classified. This is discussed in more depth
in “Choose the right measure”, below.

Avoid the temptation to use fancy or more-complex visual types just to make your report look more
impressive. What you want is the most-simple option for conveying your story. Horizontal bar charts and
simple line charts can convey information quickly. They are familiar and comfortable and most readers
can interpret them easily. An added advantage is that most people read left-to-right and top-to-bottom
and these two chart types can therefore be scanned and comprehended quickly.

Does your visual require scrolling to tell the story? Avoid scrolling if you can. Try applying filters and
making use of hierarchies/drilldown, and if those don’t eliminate the scrollbar, consider choosing a
different visual type. If you can’t escape scrolling, horizontal scrolling is tolerated better than vertical
scrolling.

Even when you choose the absolutely-best visual for the story, you might still need help telling the story.
That’s where labels, titles, menus, color, and size come in. We’ll discuss these design elements later in the
section titled “Design elements”.

Choose the right measure

Is the story your visual telling compelling? Does it matter? Don’t build visuals for the sake of building
visuals. Maybe you thought the data would tell an interesting story, but it doesn’t. Don’t be afraid to start
over and look for a more-interesting story. Or, maybe the story is there but it needs to be measured in a
different way.

For example, say you want to measure the success of your sales managers. What measure would you use
to do this? Would you measure that best by looking at total sales or total profit, growth over previous
year or performance against a target goal? Salesperson Sally might have the largest profit, and if you
showed total profit by salesperson in a bar chart, she would look like a rockstar compared to the other
salespeople. But if Sally has a high cost of sales (travel expenses, shipping costs, manufacturing costs,
etc.), simply looking at sales doesn’t tell the best story.

Reflect reality/don’t distort reality

It’s possible to build a visual that distorts the truth. There’s a website where data enthusiasts share “bad”
visuals. And the common theme in the comments is disappointment in the company that created and
distributed that visual. It sends the message that they can’t be trusted.

So create visuals that don’t intentionally distort reality and that aren’t manipulated to tell the story you
want them to tell. Here is an example:

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 19/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 22: Distorted reality chart

In this example, it appears as if there is a big difference between the 4 companies, and that CorpB is way
more successful than the other 3. But notice that the X axis doesn’t start at zero and that the differences
between the companies is likely within the margin of error. Here’s the same data with an X axis that does
start at zero.

Figure 23: Realistic chart

Readers expect and often assume the X axis is starting at zero. If you decide to not start at zero, do so in
a way that doesn’t distort the results and consider adding a visual cue or text box to point out the
deviation from the norm.

Design elements

Once you’ve selected a type and measure and created the visual, it’s time to fine-tune the display for
maximum effectiveness. This section covers:

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 20/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Layout, space, and size

Text elements: labels, annotations, menus, titles
Sorting
Visual interaction
Color

Tweaking visuals for best use of space

If you’re trying to fit multiple charts into a report, maximizing your data-ink ratio will help make the story
in your data stand out. As mentioned above, Edward Tufte coined ‘data-ink’ ratio: the goal is remove as
many marks from a chart as possible without impairing a reader’s ability to interpret the data.

In the first set of charts below, there are redundant axis labels (Jan 2014, Apr 2014 etc.) and titles (“by
Date”). The titles for each chart also require dedicated horizontal space across each chart. By removing
the chart titles and turning on individual axis labels we remove some ink and have better use of the
overall space. We can remove the axis labels for the top two charts to further reduce ink and use more of
the space for data.

If there were particular time periods that you wanted to call out, you could draw lines or rectangles
behind all the charts to help draw the eye up and down to aid comparisons.

Figure 24: Before

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 21/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 25: After

To turn axis titles on and off

Select the visual to make it active and open the Formatting pane. Expand the options for the X-axis or Y-
axis and drag the slider for Title on or off.

Figure 26: Turn axis titles on and off

To turn axis labels on and off

Select the visual to make it active and open the Formatting pane. Next to X-Axis and Y-Axis are sliders.
Drag the slider to turn axis labels on or off.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 22/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 27: Turn axis labels on and off

 Tip

One scenario where you might turn Y-axis labels off would be if you had Data labels turned on.

To remove visual titles

Select the visual to make it active and open the Formatting pane. Set the slider for Title to Off.

Figure 28: Remove titles from visuals

Consider how your readers will be viewing the report and ensure your visuals and text are large enough
and dark enough to be read. If you have a proportionally-larger visual on the page, readers may assume
it’s the most important. Put enough space between the visuals that your report doesn’t look cluttered
and confusing. Align your visuals to help direct the eyes of your readers.

To resize a visual

Select the visual to make it active. Grab and drag one of the handles to adjust the size.

Figure 29: Resize visual

To move a visual

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 23/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Select the visual to make it active. Select and hold the gripper bar at the top middle of the visual and
drag the visual to its new location.

Figure 30: Move a visual

Titles and labels that are part of the visualizations

Ensure titles and labels are readable and self-explanatory. Text in titles and labels must be an optimal size
with colors that stand out (such as black instead of the default grey). Remember our style guide (see
"Text" above)? Limit the number of colors and sizes -- too many different font sizes and colors make the
page look busy and confusing. Consider using the same font color and size for the title of all visuals on a
report page and choose the same alignment for all titles on a report page.

The formatting pane

For each of the formatting adjustments listed below, select the paint roller icon to open the Formatting
pane.

Figure 31: Open the Formatting pane

Then select the visual element to adjust and make sure it is set to On. Examples of visual elements are: X-
Axis, Y-Axis, Title, Data labels, and Legend. The example below shows the Title element.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 24/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 32: Format a visual title

Set the text size

Text size can be adjusted for titles and data labels, but not for X or Y axes or legends. For data labels
specifically, play with the Display units and number of Decimal Places until you find the optimal level of
detail for displaying on your report.

Set the text alignment

The choices for title alignment are left, right, and center. Choose one and apply that same setting to all
visuals on the page.

Set the text position

Text position can be adjusted for some Y axes and for the legend. Whichever you choose, do the same
for the other Y axes and any other legend on the page.

Set the title and label length

Adjust the length of titles, axes titles, data labels, and legends. If you decide to display any of these
elements, adjusting the length (along with text size) ensures that nothing is truncated. For Title and
Legend, the setting is Title Text and this is where you type in the actual title that will appear on the
visual. For X-Axis and Y-Axis, the setting is Style and you select from a dropdown. For Data labels, the
settings are Display and Decimal. Use the Display dropdown to select the units of measurement:
millions, thousands, none, auto, etc. Use the Decimal field to tell Power BI how many decimal places to
display.

Set the text color

Text color can be adjusted for titles, axes, and data labels.

Titles and labels that are not part of the visualizations

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 25/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Earlier in this paper we discussed adding text boxes to report pages. Sometimes the titles on
visualizations aren’t enough to tell the story. Add text boxes to communicate additional information to
the readers of your reports.
To keep your report page from looking too confusing or too busy, be consistent in your use of text box
fonts, sizes, colors, and alignment. To make an adjustment to the text in a text box, select the text box to
reveal the formatting menu.

Figure 33: Format the font used in a text box

Sorting

A really simple opportunity to provide faster insight is to set the sorting of visuals. For example, sorting
bar charts in descending or ascending order based on the value in the bars enables you to quickly show
significant incremental information without using more real estate.

To sort a chart, select the ellipses (…) in the top right of the chart, select Sort and choose the field you
want to sort by and the direction. For more information, see Change how a visual is sorted.

Chart interaction and interplay

One of the most compelling feature of Power BI is the ability to edit the way charts interact with each
other. By default, charts are cross-highlighted: when you select a data point, the related data of other
charts light up and the unrelated data dims. You can override this behavior to use any chart as a true
filter which saves you real estate on your page. To do this, select Visual Interactions from the menubar.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 26/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 34: Visual interactions

Then, for each visual on the page, decide whether you want the selected visual to filter, highlight, or do
nothing. Not all visuals can be highlighted, and for those the highlight control won't be available. For
more information, see Visual interactions in Power BI.

 Tip

For readers who’re new to Power BI, this ability to click and interact with reports may not be instantly
obvious. Add text boxes to help them understand what they can click on to find more insights.

The use of color in visuals

Earlier in this paper we talked about the importance of having a plan for how you’re going to use color
across a report. This section will have some overlap but primarily applies to how you use color in
individual visuals. And the same principles apply: use color to tie the report together, add emphasis to
important data, and to improve the reader’s comprehension of the visual. Too many different colors is
distracting and makes it difficult for the reader to know where to look. Don’t sacrifice comprehension for
beauty. Only add color if it improves comprehension.

 Tip

Know your audience and any inherent color rules. For example, in the United States, green typically
means “good” and red typically means “not good”.

This topic is broken down to cover:

1. Data color
2. Data label color
3. Color for categorical values
4. Color for numerical values

Use colors to highlight interesting data

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 27/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

The simplest way to use color is by changing one or more data point’s color to call attention to it. In this
example, the color changes when the Olympic games moved from a 4-year cycle to a 2-year cycle of
alternating Summer and Winter games.

Figure 35: Use color to tell a story

You can change data point colors from the Data colors tab in the formatting pane. To customize each
data point individually, make sure Show all is set to On.

Figure 36: Set data point colors

７ Note

Power BI applies a default theme to your report visuals. The theme colors have been chosen to
provide variety and contrast. To divert from the default theme palette, select Custom color.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 28/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 37: Choose a custom color

In Power BI Desktop, you can even highlight outliers or a section of a line by using a second series:

Figure 38: Using Desktop to plot outliers

Here, values in the ‘Outliers’ series only exist where the average August temperature drops below 60.
This was done by creating a DAX calculated column using this formula:

Outliers = if(Editions[Temp]<60, Editions[Temp], BLANK())

In our example, there were 3 outliers: 1952, 1956, and 2000.

Colors for labels and titles

As you explore all the formatting options available to you, you’ll find many different places to add color
to titles and legends. For example, you can change the color of data labels and axes titles. Proceed with
caution. Generally, you want to use a single color for all visual titles. As with all the guidelines in this
paper, there are always situations and reasons to “break the rules”, but if you do decide to break the
rules, do it for a good reason.

Colors for categorical values

Charts with a series typically have a categorical value in the legend. For example, each color in the legend
below represents a different category of Country/Region.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 29/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 39: Default colors applied

The colors Power BI uses by default were chosen to provide a good color separation between categorical
values so they are easy to distinguish. Sometimes people change these colors to match their corporate
scheme etc. but it can lead to problems.

Figure 40: Color applied as hues of a single color

By sticking to a single hue and varying the intensity of the color, this visual has introduced a false sense
of ordering between the categories. It implies the darker bubbles are higher or lower on some scale than
the lighter hues. Other than alphabetical, there’s normally no inherent order in this sort of categorical
value. To change the default colors, open the Formatting pane, and select Data colors.

Colors for numerical values

For fields that do have some inherent order and numerical value, you can also color data points by the
value. This can be helpful to show the spread of values across the data, and also allow for two variables
to be shown on a single chart. For example this chart makes it clear that although China has the highest
medal count, Japan and Thailand have participated in more Olympic games.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 30/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 41: Color data points by the value

To create this chart, add a value to the Color saturation field and then adjust those colors in the
Formatting pane.

Figure 42: Add a color saturation field

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 31/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 43: Adjust the colors used for saturation

Color can also be used to emphasize variance around a central value. For example, coloring positive
values green and negative values red. Be aware of cultural differences when assigning colors to positive
or negative values; not all cultures use red for bad and green for good!

Figure 44: Color to emphasize variance around central value  

Principles of visual design – applied to example report page

Now let’s take the visual principles discussed above and apply them to our sample report.

Before

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 32/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 45: Our example report (before)

After

Figure 46: Our example report (after)

What did we do?

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 33/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

1. Slicer: removed blanks from the slicers by adding a page level filter and selecting only gold, silver,
bronze. Changed Selection Controls to Off for Single Select and Select All.
2. Bubble: there are so many items in the legend that they scroll off the screen. Removed the legend
and turned on Category labels instead. Customers can hover over the bubbles to see the details.
Shortened the title and removed “by countryregion” since that seems self-evident. Turned axes
labels On for both to make the chart easier to understand.
3. Filled map: changed the Data colors to make it stand out more. Turned Diverging on and set the
Minimum to pink and the Maximum to red.
4. Tree map: removed filter which was set for only USA. Set the Data labels to 1 decimal place. The
visual was using the Class field which isn’t very useful since it will almost always be 33%
(Gold/Silver/Bronze). Selected a different more-interesting field, gender. Changed Aquatics to blue
and Athletics to grey for design.
5. Top bar chart: shortened the title, removed data labels, turned legend title off. Changed word order
of title to match the chart below.
6. Bottom bar chart: sorted by year ascending to match chart above. Changed colors to match class.
Changed title. Turned off legend for more space for data. Turned on data labels which won't show
up in the report (because the visual is too small for the labels to be readable) but will show when
the visual is opened in Focus mode. Learn about Focus mode. Added Count of Event (Distinct) to
Tooltips so now when you hover over a stacked column, the tooltips also tell you how many events
were contested that year.
7. Visual Interactions: turned off interactions for both cards since I always want them to show total
games and sports.

Visual types and best practices

Power BI provides many visual types natively. To these, add the custom visuals available from Microsoft
and from the Power BI community and total visual options become too numerous to document here. But
let’s look at some of the most-used native visual types.

Line charts

Line charts are a powerful way to look at data over time. Looking at data in tables doesn’t really take
advantage of the speed in which our eyes spot peaks, valleys, cycles, and patterns.
The example below shows the trends in the number of medals awarded and the number of athletes
winning those medals.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 34/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 47: Line charts

Best practices

When people look at line charts, the first thing they see is the shape of the curve. This means that
you need to have an x-axis that makes the curve meaningful such a time or distribution categories.
If you put categorical fields like product or geography on the x-axis, the line chart will not be
interesting as the shape of the curve would provide no meaningful information.

If you choose to place multiple charts above and below each other like this, to make it easier to
compare across series, line up the X-axis. Use filters to make sure that the same range of values is
shown. For example, if you’re looking at date ranges, ensure they are the same date ranges. For
example, 1896 to 2012 on both charts.

Make full use of the space. If it makes sense for your data, set the start and end points for the Y-
axis to eliminate empty space at the top and bottom of your chart and to focus in on the actual
data points. To do this, select the paint roller icon to open the Formatting pane. Expand the Y-Axis
area and set the Start and End points.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 35/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 48: Set the start and end points

Another reason to explicitly set the Start and End points is if you’re comparing two or more charts
on the same page using the same Y-axis field. For example, if you’re looking at cumulative event
counts, and the United Kingdom has counts that range from 1 to 70 and Australia has counts that
range from 1 to 12, the 2 line charts will display very different Y-axes (Figure x). This makes it
difficult to compare at a glance. Instead, set the charts to use the same Y-axis range (Figure x).

Figure 49:Line charts with different y-axes

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 36/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 50:Line charts with matching y-axes

For more information, see:

Customize the X and Y axes

Line charts and irregular intervals
Line charts 101

Bar/Column Charts

If line charts are the standard for looking at data over time, bar charts are the standard for looking at a
specific value across different categories. If you sort the bars based on the number, you will instantly see
the top values and distribution. Horizontal bar charts work well with long-ish labels.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 37/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 51: Horizontal bar chart

Best practices

Display data labels for values. This makes it easier to identify specific values. To do this, open the
Formatting pane, and set Data labels to On.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 38/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 52: Turn on data labels

The bar chart above is really useful to compare one measure against many at a single point in
time. While the line chart above showed us the trend over time, the bar chart shows us the trend
for a single category at a specific point in time. At a glance, our bar chart shows us Spain has one of
the worst unemployment rates in the world, at 25%.

When an entire Bar/Column chart doesn’t fit into the allotted space, Power BI adds scrollbars. When
possible, and if it makes sense, structure the visual and report to show the entire chart so the
reader gets an overview of the entire distribution. Unfortunately this is not possible in our example
given the significant number of countries around the world.

One way to limit the values included is to use a filter. For example, add a Visual level filter that
shows the country only if unemployment rate is above 20%.

Bar/Column charts can be drilled down (and back up again). This is a great way to pack more
information into a visual without taking up more real estate. The example below has a hierarchy for
Regions > Countries. Double-clicking a region bar drills down to the countries that make up that
region. For more information on drill, see Drill down in a visualization.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 39/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 53: Drill down

For more details on Bar and Column charts:

Bar charts 101

Data Visualization Catalogue: Bar Chart
Data Visualization Catalogue: Multi-set Bar Chart

Stacked Bar/Column Charts

Add another dimension to your bar/column charts by stacking different categories within the bar or
column. Now the chart conveys information about one overall trend (based on height/length) but also
shows the influence of the categories on that trend. The chart below shows the overall growth of Top
soccer team revenue above 6 billion in 2014.

Figure 54: Stacked column chart

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 40/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

This stacked column chart shows us that total revenue is growing over time and that the Commercial and
the Broadcasting categories are increasing steadily over time – contributing to overall revenue increase.
But this chart doesn’t make it easy to compare the impact each of the 3 categories has on each other.
For example, how does the growth of Commercial compare to the growth of Broadcasting or Match
Day? A better choice for this data, or a companion visual for this data, would be a line chart.

Figure 55: Convert to a line chart

In this line chart it is easier to see how commercial revenue has grown the most followed by broadcast
and match day.

Best practices

As with columns/bars, you have the option of horizontal or vertical display. Horizontal is a better
choice if you have long labels and vertical if you have time series data.

Avoid stacked Bar/Column charts if you want to show trends and other patterns of change over
time. Other charts, like Line charts, do a much better job.

You can also have the distribution based on total volume or as a % of total.

As Few noted it is difficult to compare the segments of a stacked bar. If the segments were arranged
side-by-side and all grew upwards from the same baseline, it would be easy to compare their heights,
but when stacked upon one another, the task becomes hard. Plus, although it’s fairly easy to see how
(revenue) changed from month to month it is quite difficult to see how (revenue) in the other
(categories) changed.

100% Stacked charts are a good choice when using percentages that add up to 100. In the example
below, we see the category distribution by team. The percentages are relative and allows us to, at a
glance, see patterns. For example, Everton’s revenue comes primarily from Broadcasting (over 70%)
https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 41/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

while PSG only derives 20% of its revenue from Broadcasting. The choice of a horizontal display
makes it easier to fit the team labels and to see the impact of revenue type.

Figure 56: Horizontal stacked chart

For more information on stacked charts:

Data Visualization Catalogue: Stacked bar graphs

When are 100% stacked bar graphs useful?

Combo Bar/Column Charts

In Power BI, you can combine column and line charts into a combo chart. The choices are: Line and
Stacked Column chart and Line and Clustered Column chart. Save valuable canvas space by combining
two separate visuals into one.

The two screenshots below show a before and after. The first page has two separate visuals: a Column
chart showing population over time and a Line chart showing GDP over time. These charts are a good
candidate for a Combo chart because they have the same X-Axis (year) and values (2002 through 2012).
Why not combine them to compare these 2 trends on a single visual? Combining these 2 charts lets you
make a quicker comparison of the data.

The new report page has a single visual: a line and stacked column chart. We could’ve just as easily
created a line and clustered column chart. It’s now easier to look for a relationship between the two
trends. We can see that up until 2008, population and GDP followed a similar trend. But starting in 2009,
as population growth flattened, GDP was more volatile.
https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 42/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 57: As two separate charts

Figure 58: As a single combo chart

Best practices

Combo charts work best when both visuals have at least one axis in common.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 43/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Watch your axes! Is your Combo chart easy to read and interpret? Or does it use dissimilar ranges and
values? For example, if the scale of the column chart’s Y-Axis is much smaller than the scale of the line
chart’s Y-Axis, your combo chart won’t be meaningful. For example, notice the third line (aqua color) way
down at the bottom.

Figure 59: An unsuccessful line chart

So too, your combo chart won’t be meaningful if your column chart and line chart use 2 different
measures and you don’t create dual axes. For example, dollars versus percent. Be sure to include both
axes to help the reader understand the chart and consider adding axes labels as well.

To do this, open the Formatting pane, expand Y-Axis and set Show Secondary to On (if it isn’t already
on). This setting is sometimes difficult to find; expand Y-Axis (Column) and scroll down until you see
Show secondary. Also, set the Y-Axis (Column) Title to On and set the Y-Axis (Line) Title to On.

Figure 60: Show secondary axis

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 44/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 61: Create a combo chart instead

Take advantage of dual axes. It’s a great way to compare multiple measures with different value
ranges. And it’s a great way to illustrate the correlation between two measures in one visual.

For more information:

Combo chart in Power BI

The danger of dual-scaled Axes in visuals

Scatter Chart

Sometimes we have many variables that we want to see together, and a scatter chart can be a very useful
way to get an overall picture. Scatter charts display relationships between 2 (Scatter) or 3 (Bubble)
quantitative measures. A scatter chart always has two value axes to show one set of numerical data along
a horizontal axis and another set of numerical values along a vertical axis. The chart displays points at the
intersection of an x and y numerical value, combining these values into single data points. These data
points may be distributed evenly or unevenly across the horizontal axis, depending on the data.

A bubble chart replaces the data points with bubbles, with the bubble size representing an additional
dimension of the data.

The bubble chart below looks at South America and compares GDP per capita (Y-Axis) sum of GDP (X-
Axis) and population by South American country. The size of the bubbles represents total population for
that country. Brazil has the largest population (bubble size) and the largest share of South America’s GDP

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 45/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

(it is farthest along on the X-Axis). But notice that GDP per capita for Uruguay, Chile, and Argentina is
higher than Brazil (farther up on the Y-Axis).

Figure 62: South America GDP and population as a bubble chart

If you add a play axis, you can pretend you are Hans Rosling and tell the story over time
(https://www.youtube.com/watch?v=PbaDBJWCeD4). To add a play axis, drag a datetime field into the
Play Axis well.

Best practices

Scatter and Bubble charts are great storytellers. But they are not as useful when trying to explore
data. This is what Stephen Few points out in the paragraph below The strength of this approach is
when it’s used to tell a story. When Rosling narrates what’s happening in the chart as the bubbles
move around and change in value, pointing to what he wants us to see, the information comes alive.
Animated bubble charts, however, are much less effective for exploring and making sense of data on
our own. I doubt that Rosling uses this method to discover the stories, but only to tell them once
they’re known. We can’t attend more than one bubble at once as they’re moving around, so we’re
forced to run the animation over and over to try to get a sense of what’s going on. We can add trails
to selected bubbles, which make it possible to review the full path these bubble have taken, but if
trails are used for more than a few bubbles the chart will quickly become too cluttered. Essentially,
what I’m pointing out is that this is not the best way to display this information for exploration and
analysis.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 46/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Add X and Y axes labels to help tell the story. Especially with bubble charts, there are many
components at play and labels help readers understand the visual.
Add data labels to make the visual easier to interpret. Especially with bubble charts, when you have
many items in the Legend, it may be difficult to distinguish between similar colors. In the visual
above, the legend colors for Suriname, Columbia, and Ecuador are very similar.
Did you create a scatter chart and see only one data point that aggregates all the values on the X
and Y axes? Or, your chart aggregates all the values along a single horizontal or vertical line? To fix
this, add a field to the Details area to tell Power BI how to group the values. The field must be
unique for each point you want to plot. For help, refer to the Power BI scatter and bubble chart
tutorial.

Tree Map Charts

Tree maps can be very useful for giving a good overview of the relative size of different components that
make up a whole -- especially when you can group them by categories. Any time I try to understand a
new business, having a tree map of the main components can be very useful in knowing the overall
distribution.

In the first chart below, you can see right away that Brazil makes up approximately half of South
America’s GDP and that Venezuela and Argentina are roughly the same size.

If you want to have broader context and still have an idea of the impact of the top contributing
countries, you can create visual hierarchies with category members (countries) nested inside regions. The
second tree map gives us an idea, first and foremost, of the relative size of the regions and then, within
each region, we can see which individual countries contribute the most. We see that there are three
massive regions (Europe, Asia & North America) and within those we can easily see the top
countries/regions.

The main limitation of a tree map is the limited ability to compare the different rectangles beyond the
top ones. It is a good chart for an overview but column and bar chart are probably a better choice to
have more precise idea of the relative size of different components.   For example, the first tree map
gives a broad indication of the order of the GDP size, but it’s hard to identify specific differences
between countries, particularly the smaller unlabeled boxes. For this data, where a single grouping is
compared, a bar or column chart might be a better choice.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 47/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 63: South America GDP comparison as a tree map

Here we’ve added another level of data, region, and we can see the overall contribution to GDP by
regions, as well as the relative impact within the regions. Beware that doing this with non-summative
measure (such as averages) that the sum of the details might not represent the actual value at the
aggregate level.

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 48/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Figure 64: GDP by region and country as a tree map

For more information on tree maps, feel free to click on links below.

Treemaps overview
Data Visualization Catalogue: Tree maps

Other charts

Pie or Donut Charts

In general, bar/column/line charts will serve most purposes. It’s well understood that pie and donut
charts are difficult for humans to interpret correctly, and in fact can often distort data. Avoid them where
possible. Stephen Few has an excellent write up on the history and dangers in [Save the Pies for Dessert]
(www.percetualedge.com/articles/08-21-07.pdf

He does explain the one time where pie charts can be useful, when comparing part-to-whole
relationships. But even this is rarely significantly-better than, say, a 100% stacked bar chart.

Another fun article (and animation) about pie charts can be found on the Darkhorse Analytics site.

Radial Gauges & KPIs

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 49/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

Radial gauges seem like a good visual for indicating performance against a target, and they are very
popular in executive dashboards. However, they suffer in two main ways. As with pie charts, it’s difficult
to interpret the angle of the shaded area compared to the full 180 degree arc or target line. It also uses a
lot of space to show a single metric.

A good alternative is a simple KPI visual

KPIs show the value, status, goal, variance from the goal and trend in the same amount of space. The
green coloring turns red if the target isn’t being met and can be yellow if some intermediate target is hit.
It’s much simpler to read and interpret than the gauge.

For more information, see:

Radial gauge charts in Power BI

KPIs in Power BI

Conclusion
Now it’s time for you to put these best practices to the test. Keep in touch and share your own best
practices. Don’t agree with our recommendations or found a great reason to “break the rules?” We’d love
to hear about those as well.

Book recommendations

There are many good books available today to help teams bone up on visual design techniques. Stephen
Few’s Information Dashboard Design book is a must-read. He delves into greater detail in two other
books, Show Me the Numbers and Now You See It. Few and others have drawn inspiration from Edward R.
Tufte, whose book The Visual Display of Quantitative Information is considered a classic in the field. Tufte
has also written Visual Explanations, Envisioning Information, and Beautiful Evidence. Andy Kirk’s new
book Data Visualization: A Handbook for Data Driven Design is another great option. Some other authors
https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 50/51
1/13/2019 Best design practices for reports and visuals (whitepaper) - Power BI | Microsoft Docs

who have been recommended are: Lachlan James, William McKnight, and Boris Evelson (Forrester),
Darkhorse Analytics.

More questions? Try the Power BI Community

https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-best-practices 51/51