Designing your charts for PowerPoint Presentations

December 21, 2020

Take a look at this chart below. Let’s picture it in two different scenarios.

Scenario one

It’s Monday morning. You grab your cup of coffee, sit down at your computer and start going through your emails. Your colleague has sent you an interim report on findings from their current research project. You scroll through, read the report and take time to understand and interpret the chart. “Excellent work”, you say. You go about your business.

Scenario two

Your same colleague has been invited to present the findings from the interim report. Being 2020, let’s say this is on Zoom. They have 15 minutes to get through about 20 slides, which leaves about 45 seconds per slide. One of the slides is the same chart. The chart appears. They keep talking, and you start trying to read, interpret and make comparisons between the categories. However, your concentration is broken because the presenter is still talking and what they are talking about is not where you are on the chart. Next minute, the slide changes. “I’m confused”, you say.

My point is when designing charts for your PowerPoint presentations, there are different considerations to designing for reports or infographics. With reports and infographics, you have a lot more time to read and digest the information. For presentations, we need to reduce the cognitive load and simplify the chart as best we can to illustrate the point.

So, what can we do?

The chart above was one that came from one of my awesome clients. The horizontal grouped bar is great and does everything a bar chart needs to. She has added data labels and removed all the background noise like the gridlines. But it still takes a bit of work to go back and forth from the key down the bottom and match up with the categories and how they perform against each other. Time is not usually a luxury we have in PowerPoint presentations.

AFTER: Small multiple horizontal bar chart

This is my redesign as small multiple horizontal bars.

What have I done?

I’ve made two horizontal bars sitting next to each other with the exact same dimensions (scale, height and width). The bars are bit thicker too. I’ve kept with my client’s chosen colours for male and female (and not conformed to the old pink and blue stereotypes) and added an icon next to the heading to aid in comprehension. The first category (female), I have listed from biggest to smallest. It makes the chart easier to read. I’ve then matched the categories for the male column for easier comparisons. Like the original, I have the data labels in there but have made them that little bit bigger. There is a big difference in the number of female and male participants, so I have kept those numbers in but subtly at the top.

A few easy changes and the chart is a whole lot easier to digest.

Let’s look at another slide before and after.

I can hear you asking, “Why have you been boring and done it in the exact same way”? Good question!

Remember: time is not our friend in a PowerPoint presentation. Even if the slide was up on screen for a while, we still want our audience to process the information as quickly as possible and then focus their attention back to us as the presenter. So, keeping the same type of chart and the same layout allows our audience a quicker processing time as they are already familiar with the chart and how to read it.

Other key things to remember when designing charts for PowerPoint presentations:

Remove unnecessary clutter from your charts, like gridlines, background colour, and borders.
Always label data directly to reduce any ambiguity. It’s harder for our eyes to scroll from the end of the bar to the X or Y axis to see the exact value.
Use the graphic design principle of hierarchy to order your chart. In Western culture we read from top left to bottom right, so lead your audience through with a large title on the top left.
Add icons where you can to aid in readability. Remember PowerPoint now has its own icons.
Use the real estate on your slide! Use the space you have been given. Don’t cram it into one corner.
Use data storytelling to focus your audience’s attention on a certain part of your chart and tell the story behind it.

Happy designing!

Let’s talk small multiples

September 29, 2020

Check out this line chart below. As the title suggests it’s looking at AFL club membership from 2016 to 2020. It’s extremely busy with 18 different lines and colours that are awfully similar. It’s hard to gain any insight from the chart (besides that fact that it’s quite busy), let alone focus on any single data series. What if I asked you to tell me how club membership was affected by COVID-19 from 2019 to 2020? Which clubs did well? Which ones lost members? What was the overall trend? It’s virtually impossible which makes the chart pretty useless.

Luckily, we can redo this chart using small multiples instead.

What are small multiples?

The term was popularised by Edward Tufte and as he describes “resemble the frames of a movie: a series of graphics, showing the same combination of variables, indexed by changes in another variable.” They are also referred to as trellis, lattice, grid or panel charts. They are an efficient way to break down a busy chart that allows your audience to walk through one category at a time and make comparisons.

So let’s redo this chart using small multiple line charts.

What have I done?

I’ve created an individual chart for each team showing membership numbers over the last five years. Each chart is to the same scale, which is incredibly important so we don’t mislead our audience. I have decided to colour code the chart with orange used for membership decline over the previous season and blue to indicate membership growth. I have then finished this off with a descriptive title and subtitle at the top which turns my chart from an exploratory chart to an explanatory one (i.e. I’m telling you what my take away message is). Lastly, I have listed my source down the bottom. Has this redesign made it easier for you to understand and interpret the data?

Let’s try another design using small multiples area charts. This time I will incorporate data storytelling.

I’ve pushed everything to the background using grey. I’ve then decided I’ll focus on membership decline once again and highlight that using a bold red colour (a little alarmist, but intentional given the circumstances). I’ve used a complete sentence to highlight my story and colour code it to the chart. Lastly, of course, my source at the bottom.

Depending on where I am using the chart (a report or PowerPoint presentation), I may want to give context that Melbourne was hardest hit by a second wave of coronavirus and teams were moved to hubs predominately in Queensland. Additionally, no Melburnians were able to attend live games within Victoria in 2020.

How have I done these small multiples?

Great news, it’s all in Excel! Highlight and create your first chart with only the one category.

Once you have formatted it the way you want it, right click and save as template. Then copy and paste your chart however many times you need to. In my case, it was 17 more times. Lastly right click on each chart and choose “select data” to populate the new data.

Small multiples can be used for just about any type of chart. My blog on displaying Likert scale data uses a few more examples.

If you need some further inspiration that’s not football related… (I had to, it’s September!), check out these links.

Financial Times: Coronavirus tracker

McKinsey & Company: Tracking US unemployment through the COVID-19 crisis

Happy designing!

Transforming your charts using data storytelling

August 30, 2020

Our team has been working in partnership with the Doherty Institute on a longitudinal, nation-wide survey. The aim is to gain an understanding how Australians are thinking, feeling and behaving in relation to the COVID-19 pandemic and the distancing measures in place. A series of rapid reports are being produced to help decision makers in real time and you can read the full reports here. I was lucky enough to work on the design and layout of the second report.

There were a number of charts produced through the analysis by the very talented data analysts on the team. One such chart is below. Can we take a moment to recognise the great colours used! The rapid turnaround of this report limited the time available for design, but now that I have had some breathing space I thought I would show you some alternative ways the information in this chart could be displayed.

There is a lot happening in the chart and a lot to take in. We can see from the outset that the large chunks of green indicate that a lot of people have changed their behaviour in relation to the pandemic. The green and the purple have a common baseline which makes it easier for us to compare lengths, but the blue category caught in the middle is a little more difficult.

When I was formatting the report, I thought this would be an excellent example of how we can transform the chart to incorporate data storytelling.

What is data storytelling?

Data storytelling isn’t about counting how many breaths the big bad wolf took to blow down the three little pigs house and charting that. Or using statistical analysis to find where exactly the green sheep is, and where it might go next.

Data storytelling is finding a story behind your data and chart, and explaining that to your audience. It’s about turning your chart from an exploratory one, where you let the audience find their own meaning, to an explanatory one where, you tell them what’s important and what you would like them to take away from the chart. Ann K Emery and Cole Nussbaumer Knaflic are the real leaders in this space.

I redesigned this chart twice , then transformed each from a traditional chart into a data storytelling chart. The good news, it was 100% done in Excel!

Chart 1 – 100% stacked bar chart.

Sticking with the original concept, I used a 100% stacked bar to show behaviour change. I made one small modification which totally transforms the chart, that is ordering the first category (yes) from largest to smallest.

100% stacked bar showing behaviour change

I’ve included percentages for the yes category to give some reference. The chart appears a little more ordered now. But it’s still an exploratory chart. I leave the audience to look through and find what’s important.

Let’s take it one step further now and make it into a storytelling chart.

Using data storytelling to show our audience what is important

What did I do?

I greyed everything out.
I decided on the story I’m going to tell. In this case, my story was the majority of people did change their behaviours to some degree and thus helped protect themselves and others from the virus.
I drew attention to this by using my action colour teal to highlight the ‘yes’ category.
I wrote a full sentence heading that describes that point.
I added an extra element by placing annotations next to three bars in the chart that I thought were interesting.

The storytelling part of your chart also depends on the audience you are presenting it to. Using knowledge translation, we can identify our target audience and what aspects of the chart may be of most interest or useful to them. For instance, say I was applying for a research grant to explore whether the rate of serious illness diagnoses have decreased because fewer people have presenting to hospital for check-ups during the pandemic. I could use the above chart to show that and redesign it like this:

Storytelling chart highlighting a different point.

Again, I’ve pushed everything to the back and highlighted my story using colour, a completed descriptive heading and an annotation.

Chart 2 – Small multiple bar chart.

The second redesign uses small multiple bar charts. To help the reader, I have categorised the behaviours into four different groups; travel, safety, work and study, and social. Each group has their own colour. This breaks up the chart into smaller, more digestible chunks and makes it easier for our audience to walk their way through.

As I’ve mentioned before, people have a much easier time comparing lengths when reading charts, especially when they have a common baseline, hence my decision to go with small multiple bars.

Here is our traditional chart:

The chart is great but again, I have left it to the audience to find meaning. So, let’s tell them exactly what’s important:

Small multiple bar chart incorporating data storytelling

What did I do?

I greyed everything out.
I decided on the story I’m going to tell. Again, I wanted to focus on the behaviour change (‘yes’ category), especially the two highest categories; ‘applying social distancing rules’ and ‘avoiding crowded areas’.
I drew attention to this by leaving the yes category in colour but reducing the opacity to 50%. I left ‘applying social distancing rules’ and ‘avoiding crowded areas’ at 100% opacity, making them a bit brighter to the reader’s eye.
I wrote a full sentence heading that describes that point and used colour to link it to the chart.
This time I decided against an annotation as the chart was already quite busy.

Easy done! Your audience can’t miss the meaning now.

Using these 4-5 steps, you can transform your traditional chart into a storytelling chart making it super easy for your audience to read and interpret.

Happy designing!

7 ways to display qualitative data

July 30, 2020

Without doubt, one of the most common questions I hear in my work is “Alana, how do we display qualitative data?” The short answer is there is no hard and fast rule on this. Unlike quantitative data, qualitative data really needs you to think a little bit more abstractly. You need to ask yourself, how am I going to get my data across in a way that’s engaging and meaningful for my desired audience? How do I condense and translate these important messages and themes?

A lot of the work that I do for our Child and Community Wellbeing research team is displaying qualitative data. So, I thought I’d share some examples of just how we can do this, and how simple it can be.

So, let’s start with the basics. If you are conducting qualitative research, you are most likely conducting interviews or focus groups. The richness of data that comes out of these cannot be underestimated.

This naturally leads us to our first option.

Quotes

Now, before you start yelling at your computer screen saying, “I know this already!” hear me out. We’re not just going to make our text italics and indent it slightly. No way! I actually can’t stand italic text, but that’s a story for another day. We are going to use some graphic design magic to make those quotes pop!

How do we do it?

Bold and action colour

As Stephanie Evergreen says, we will use an action colour to highlight the important words in that quote. We will also bold those words. This breaks up the text a little, and adding those visual attributes highlights the really important parts of the quote.

Large quotation marks

I’m sure you’ve seen this before. It’s not a new trick but it’s effective. If you have a text heavy report, this is a great way to break it up and make it more digestible. On a PowerPoint slide, have nothing else but the large quotation marks and the quote.

Indent and all colour

Another great one for text heavy reports is to indent the quote and colour the whole text. A vertical line on the left of the quote, which lines up with the body text makes it look neat and tidy.

I’ve added some examples of quotes I have used in my work. With all options, make sure you have enough white space around the quote to make it stand out and not look squished.

The next obvious choice for qualitative data display is Images.

Images

As the old adage goes, ‘an image is worth 1000 words’. This article about the Thai boys who were stuck in the cave in 2018 sums it up perfectly. We are visual creatures who are drawn to images that evoke our emotions.

During data collection of the Beyond Bushfires study, our team would go out and conduct one-on-one interviews with participants. Often the participants would walk them around the local area or the bush and show particular places that were special to them. Our team always had a camera in hand and took photos to capture these moments. Many of these photos were used in the final report of the Beyond Bushfires study.

Images of the Beyond Bushfires final report

In a PowerPoint presentation, replace text or a quote with a photo that showcases that text. If you didn’t take one during data collection, there are a number of online sites that you can get free photos from like pixabay.

In a report, try to have images or graphics on most pages. It breaks up the text and makes it more digestible. When I create one-page research summary infographics, I also aim to include an image, to draw people in.

Icons

Whether it’s a quote or text explaining common themes from your qualitative data, break it up with the inclusion of an icon. Again, this is a technique I commonly use in my work. John Medina tell us in Brain Rules that when people see a chunk of text, they are likely to remember 10% of that information three days later, however when people see text combined with a relevant image, they are likely to recall 65% of that information three days later.

Use icons to show common themes in your results, or to show data collection measures. Here are some examples I have used.

Word Clouds

I’ve been very vocal with my thoughts on word clouds. There are only two times I would consider them useful. The main one is for showing comparisons over two time-points. For example, before and after or 2019 vs 2020. Even then, it’s not about dumping whole transcripts of interview data into the word cloud generator. It’s useful for short answer responses so it’s easy to see common themes emerging between the two time-points.

Here is an example from our Survive and Thrive study. We used a visual mapping exercise where children were asked a series of questions about what they would do in a bushfire and who they would give information to. This was done twice, before and after the CFA training.

Word Clouds used for comparisons over time

The other time I use word clouds is during a presentation when I showcase the Poll Everywhere software. You can ask your audience a question (usually requiring a one word response) and answers start to populate on screen in front of you. Example below from my July course. For more in depth thoughts on word clouds, please refer to my previous blog ‘Are word clouds dead? The answer is yes but no’.

Colour-coded phrases

Another great example provided by Ann K Emery and published in the New York Times, is colour-coded phrases. In this example, they have gone through Donald Trump and Hilary Clinton’s Nominees’ Convention speeches and used colour to code them into negative and positive statements about America. This simple technique gives a quick overview of the tone of the transcript. I would avoid using red and green to show negative and positive associations because of accessibility issues. Instead try blue and orange.

Venn Diagram

A Venn diagram is a great way to show different sets or themes in your research and how they overlap and intersect. We, the Child and Community Wellbeing Unit, use a Venn diagram to show how our body of research and three focus areas overlap, with our main goal in the middle.

A brilliant example of a larger Venn diagram with more intersections come from David McCandless. Here he has visualised what makes a good visualisation.

What makes a good visualization by David McCandless https://www.informationisbeautiful.net/visualizations/what-makes-a-good-data-visualization/

Timeline

The last example I will present you with is a timeline. Timelines are especially useful for visualising longitudinal studies where data collection has occurred over multiple time points.

It can show results or simply the study outline, design or even initial project plan.

Below is a detailed timeline produced for the Beyond Bushfires study looking at project engagement activities.

However, timelines don’t need to be this detailed. Above is another one I included on the Survive and Thrive infographic.

There are many more ways to display qualitative data, but I won’t overload you with too much information. I will save that for another blog post. Remember to think visually and use icons and images to help illustrate your important themes and messages.

Happy designing!

Displaying Likert Scale Data. Same data, displayed 7 different ways

May 26, 2020

Working with research and evaluations teams, I’m often faced with the question of how to effectively display Likert scale data. If you are not familiar with the term, a Likert scale, named after American psychologist Rensis Likert, measures a response to a close-ended question on a rating scale of how much you agree or disagree, often with a neutral option in between. A typical five-point scale may include answers: strongly disagree, disagree, neutral, agree, strongly agree.

I regularly use a five-point Likert scale format to evaluate my Design and Data Visualisation Short Course in order to gain specific feedback on how participants rated the quality and usefulness of each session. The challenge for a lot of people is that once they have collected the detailed data, how do they best display it in an effective way that is easy for their audience to understand and interpret?

This month, I decided to give myself a challenge. Take the same data, from a five-point Likert scale, and display it as many different ways as I could. I came up with seven different options, some much better than others. Let me walk you through.

To begin with, this is the example data I have used:

Sample data displayed from the question ‘How would you rate the quality of the sessions and activities?'

I have made some specific design choices to maintain consistency throughout, including using blue to show positive ratings, grey for neutral, and orange for negative. In some charts I have used the actual value, and others have used the relative frequency (percentage).

DISPLAY 1: SMALL MULTIPLE PIE CHARTS

The world is divided on pie charts. Many data visualisation specialists will tell you how evil they are, and I tend to agree with them. They are hard to read and interpret as they rely on our audience to decode the quantitative information by judging and comparing angles and area, which is actually quite difficult to do. However, in saying that, people still love seeing the part-to-whole relationship that a pie chart offers. Whatever your feelings are toward pie charts, I wanted to include them for argument’s sake.

I have used one pie chart for each session, showing the relative frequency of responses and colour coding it to the scale above. I’ve kept the wedges in order, starting at the top and going in an anticlockwise direction. I’ve directly labelled the wedges with their percentage but left off the smaller slices. The colour scale is simple and has enough contrast between the blue and orange that we can see differences, but we still may need to go back to the key for reference.

Our general view is that session two needs a lot of work. There is a large chunk of orange there. Session three is doing quite well, no orange in sight. It’s easy to compare within each session, but not between the sessions.

DISPLAY 2: GROUP BAR CHART

If you’re like me and use google forms to collect evaluations, you’ll notice it produces a grouped bar similar to this. However, it normally colours each line with a different hue. Studies have shown that it is far easier for us to compare length and position (bar and line chart), than area and angle (pie and donut chart)1,2. On first glance at this chart, there is a lot to take in. We can see from an overall picture that again, there is more blue than orange, which is a good thing. Generally, the blue bars are quite tall, expect from session two. My eyes are also drawn to the big gap between session two and three, and it might take a while to see that there are no “poor” or “fair” responses there. Yes, we can now compare between all the sessions, but the chart may seem too overwhelming and require significantly more visual processing time, so we risk our audience skipping over it entirely.

DISPLAY 3: SMALL MULTIPLE BAR CHARTS

This display shows six individual horizontal bar charts with our Likert scale responses ordered from excellent at the top, to poor at the bottom. Again, the colour choice of blue and orange gives a quick overall view of the data and responses. In each chart, I have labelled the values directly and kept them all to scale. It is now a lot clearer to see that there are no responses for “fair” or “poor” for sessions three.

This is one of my favourite ways to display the data. We can easily see within each session how the responses vary, and if we want to compare between sessions, our eyes only need to scroll across the screen. In terms of visual hierarchy, we start with our session number (session one, for example) on the top left, then go down and read each category (starting with excellent) and follow it along to see its value (8) and how it compares to the bars below. As mentioned previously, we are much better at reading length and position and the general public already know how to understand and interpret a bar chart.

What if you want to display the data in a more modern, non-traditional way?

DISPLAY 4: WAFFLE CHARTS

Waffle charts are usually designed with 10 or 100 squares to show the relative frequency of the given data. They are another great way to show the part-to-whole relationship that people love about pie charts. You could say they are a bit of a modern twist on traditional charts you would see in programs like Excel.

The thing I like most about this display is the gradual fade of coloured squares from blue to orange. We get a good overall picture of the data and responses, but at the same time, I think it’s vital we have the categories and values directly next to each section.

DISPLAY 5: LARGE NUMBER AND TEXT

Again, thinking a little outside the box gives us the large number and text option. It’s a great way to break up a lengthy, wordy report filled with too many dot point (or bullets).

From each session, I have made a judgement call on what I want my audience to focus on. I have combined the “very good” and “excellent” categories together in sessions one, three, four, five, and activities, as I see this as a positive response. Obviously, the thing to take away from session two is that there was an overwhelming number of negative (fair or poor) responses, and that’s where our focus should lie. I have been consistent and kept the positive responses blue and negative responses orange.

DISPLAY 6: STACKED BAR

Now we’re getting to the pointy end of business. The next two charts are probably the most effective ways to display Likert scale data. A stacked bar belongs to the part-to-whole group of charts and has us displaying the percentage of responses in proportion to one another. The issue with this chart is we would normally have a common baseline on both the left hand and right hand side. A common baseline makes it easier for us to compare lengths of each bar (as they all start at the same point). In this chart, the common baseline is only on the right hand side for the “excellent” rating. Due to the nature of this data, the left hand side baseline starts with three different categories.

Like the other charts, an overall glance makes it easy to see where we are doing well, and where we need to make improvements. For ease of reading, I have labelled each percentage value directly.

There is one way we could improve this chart to make it easier to read and that brings us up to display seven.

DISPLAY 7: DIVERGING STACKED BAR

The last option and the one I would highly recommend for this type of data is the diverging stacked bar. It’s essentially the same as a stacked bar but we are categorising the data into positive and negative categories by separating them with a thick visual baseline. In the above chart I am categorising “neutral” as a negative result. We can then measure our success or failure from that baseline, which acts as a kind of target.

It’s clear, effective and is able to show a large chunk of data neatly and succinctly. Compare this to the grouped bar above or the small multiple pie charts. Which one are you able to quickly pull meaning from?

In summary, I have presented you with seven different options for displaying your Likert scale data. There is no right or wrong answer, rather options for you to consider and get inspiration from. There are also many other things you need to consider before you start designing such as:

What medium are you designing for? Is it a report, PowerPoint slide, website, or one-page handout? How much space do you have for your charts?
Who is your audience? Will a diverging stacked bar confuse them?
What colours will you use to distinguish between positive and negative categories? Are your organisation’s branding colours too similar?
What message do you want to get across to your audience?

A lot to consider, but some options above to help your thinking. Happy designing!

References

Few, S. “Save the Pies for Dessert” Visual Business Intelligence Newsletter, August 2007.
Evergreen, S. Presenting Data Effectively, 2nd Edition. Sage Publications. 2018.

Scrap the bar chart to show changes over time

November 20, 2019

I often see bar charts used to display changes over time, or “before and after” comparisons. While this is a perfectly fine chart to use, there are much more effective and creative ways to display changes between two time points. One of my favourite ways is to use small multiple slope charts.

The following vertical bar came across my desk a while back. It looks at pre and post course evaluations. It was clearly done in Excel as we can see from the lovely (hideous) colour palette it spat out. On first glance, we can see the overall differences between the red and blue categories (pre and post) but actual values are a little more difficult to determine, as there are no guiding grid lines. Our eyes also dart back and forth from the bars to the key to match the categories.

Excel bar chart showing before, after and difference bars.

The interesting part about this chart too, is that there has been a third category added which is looking at the difference between pre and post (the first two bars). The inclusion of this third bar is adding unnecessary ‘noise’ to the chart which then takes longer for the reader to process the information. It’s also a green bar, placed next to a red bar, which would make both categories look brown to someone with colour blindness.

So how can we redo this chart? I would suggest small multiple slope charts. Here is what I came up with.

So what’s a slope chart (or slope graph)? It’s a chart that shows development over two time points. We encode the data by looking at the position of the two points and the angle of the line that joins them.

Slope graphs can have multiple categories on the one chart, like the AFL and rugby ladder examples I constantly bombard you with (see below). This example is looking at ranking over two time points.

I know what you are thinking… why have I done 5 individual charts for this course evaluation when I could just add them all on the one? Well there are a few reasons.

Firstly, to make the charts easier to read, understand and digest, I would suggest adding the values directly next to each point. In some instances, you may be able to embed the category into the chart too, next to the value (Check out this NPR example on How Each State Generates Electric Power).

In our case, each of the categories/ questions are quite long. This would make them hard to squish at the side of the post data point. Also, there are four post categories that have a value of 100%. It would make it nearly impossible adding four long questions next to the one data point, not to mention, very messy.

Having them as five small multiples, makes it easier for your audience to walk through and understand each question individually.

What else have I done to make this chart simple and elegant? I’ve used the one colour for each of the five charts to maintain a level of consistency and matched it with a grey grid. As mentioned above, I’ve added the values next to each data point, and no other numbers have been included. It’s to scale, so I don’t mislead my audience and I have a descriptive heading that introduces the data below.

Below is another example I produced for a client recently, comparing two different categories (liberal fluid strategy vs restricted fluid strategy) at two times points (pre and post surgery).

I have created both of these in InDesign, however I can confidently say you could produce a slope chart in nearly any program. In PowerPoint, draw two 10cm lines and scale your data that way (e.g 65% would be 6.5cm up the 10cm line). Here is an excel tutorial from Peltier Tech Blog.

Slope charts are an easy, simple and elegant way of displaying changes over two time points. Scrap the old-fashioned bar chart and move to something much more contemporary and effective.

It's time to break up with the pie chart

September 15, 2019

My old high school recently sent out their alumni magazine which featured the VCE results of the Class of 2018. The report was lovely and well designed, but one thing in particular caught my eye. They used a donut chart to display the university courses that the students were accepted into.

The chart looked a little something like this…

I have changed the data slightly to deidentify the school.

So, what’s wrong with that you ask? Five years ago, I probably would have told you, “nothing, it looks great”, but the more I study data visualisation the more I understand why pie/donut charts are just trouble.

Pie charts rely on our audience to decode the quantitative information by judging and comparing angles and area, which is actually quite difficult to do. Studies have shown us that it is far easier for us to compare length and position (bar and line chart), than area and angle (1,2).

With a donut chart (which is just a pie chart with a hole in the middle - stay with me, don’t get hungry), we are forced to compare the arc lengths in the middle of the circle (3), which is even more difficult.

Pie charts rely on our ability to compare area and angle. Donut charts rely on our ability to compare arc lengths

All of these issues are further compounded when we start adding 3D charts to the mix, which distorts areas and angles. I have said it once and I will say it again, don’t use 3D charts!

So, let’s work through this university course donut chart and see how we could improve it. Remember, our job as data visualisation specialists is to present the information in the most effective way so our audience is able to understand and interpret the data with very little effort.

One solution, I hear you say, is to put the values directly in the donut chart. Let’s see what that looks like.

As with any chart that has the key off to the side, our eyes have to dart back and forth to try and connect the colour to the donut slice. Have you also noticed the colour choices of the donut chart? The different shades of greens and oranges make it that little bit more difficult to match up the key to the slices. So, let’s now add the labels and the values directly to the donut, which is what I normally do.

Now it’s a little easier to read, maybe a little messier too. But the big question is, are we actually looking at the donut, or just reading the labels and values? In his newsletter, “Save the pies for dessert”, Stephen Few comments:

“Why show a picture of the data if the picture can’t be decoded and doesn’t present the information more meaningfully? The answer is: You shouldn’t. Graphs are useful when a picture of the data makes meaningful relationships visible (patterns, trends, and exceptions) that could not be easily discerned from a table of the same data.”

He is right, we get no more value from having the data presented as a donut with values directly labelled than as a plain old table.

So how would I redo this? A horizontal bar. It might seem boring, but it is easy to understand and make comparisons between each category. As I mentioned before, we are much more capable of comparing length (bar) and position (line) than area and angle. Never underestimate the value of a bar chart.

I’ve ordered the data from biggest to smallest. Labelled each bar for easy understanding and got rid of the “noise” by just using the one colour. As the category labels were quite long, a horizontal bar worked better than a vertical one. I’ve also added a descriptive heading and shown the total 100% down the bottom.

Compare the before and after. Which one is easier to draw meaning from?

Pie charts might be aesthetically pleasing (when done well), but as I’ve shown you above, they are not the most efficient or effective way of displaying data. Think of it with this analogy: it’s like buying a great pair of shoes that look amazing but are horribly uncomfortable, so you never wear them. What’s the point?

It’s a hard relationship to break up. It’s taken me a lot longer than it should have. I even used a donut in last year’s annual report. But don’t feel bad, even Steve Jobs used one… in 3D!

References:

1. Few, S. “Save the Pies for Dessert” Visual Business Intelligence Newsletter, August 2007.

2. Evergreen, S. Presenting Data Effectively, 2nd Edition. Sage Publications. 2018.

3. Nussbaumer Knaflic, C. Storytelling with Data. Wiley Publications. 2015