Visualising the future we want.
Four rules for better data visualisation.
Sometimes it seems to me that the only news we hear about nature is bad news. ‘Dangerous declines,’ ‘Time is running out,’ and ‘We’re all going to die,’ are all-too-familiar headlines. The latest doomsday headlines have come courtesy of IPBES, the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, who recently published a 40-page summary on the state of nature. But it wasn’t the headlines that distressed me most. It was the sprawling, convoluted assemble of data visualisations in the report that left wondering, “what can I do to improve this?”
Realising that the summary report is an advance draft, I figured that there might be time to offer some feedback that could be used in the production of the full and final report. With that in mind, I asked my colleagues Melanie Herrmann and Dani Caso to review the visualisations with me. Melanie is a Human-Computer Interaction Researcher who aims to understand the psychology behind effective data visualisation and how we can make better, more impactful data visualisations. Dani is a Designer who combines aesthetics and functionality to communicate data in memorable, usable ways.
Together we’ve compiled some examples, explanations and edicts that can be used by anyone who wants to ensure their research is noticed, shared, and used in ways that lead to positive action for our planet. You’re about to see how much your choice of data design influences the message it communicates.
IPBES, policymakers, and difficult decisions.
IPBES is one of the foremost scientific bodies in the world. Policymakers — who need reliable information from a trustworthy source — will listen to what they have to say. However, not all policymakers are trained in data analysis, and they have many demands on their attention and time. If we want these people to take note of what IPBES is trying to tell them, we need to deliver information in the most efficient, effective way possible.
Data visualisations are a great way to communicate information. A review by SciDev.net on the reach of their own research found that well-designed visuals were more likely to be picked up than a big report. When done well, data visualisations can plant seeds in our minds that blossom into sustainable solutions for the planet. But when they’re not, data visualisations can be difficult to interpret, obscure, or at worst, outright misleading.
The data visualisations in the IPBES summary report have four main issues:
They are not pretty.
Ugly data visualisations are less memorable and people will be less inclined to use the data.
They contain too much information.
Too much detail dilutes the key insights you’re trying to highlight.
It’s all doom and gloom.
Bad news can overwhelm and scare people, making it more likely that they’ll ignore what you are telling them.
They’re hard to find and share.
If information is hard to find and share, it won’t be seen or used.
Choose Beauty, not the Beast.
Our parents told us not to judge a book by its cover. And yet, our subconscious brains still judge how good something is by how pretty it is. Even if we don’t want to admit it, we are more inclined to trust what we’re seeing or hearing if it’s delivered in an attractive way. In the same way that I’d happily listen to Henry Cavill read numbers from a phone book all day, I’d commit time to exploring and absorbing a beautiful data visualisation.
When it comes to data visualisations, beauty offers two advantages. Firstly, viewers will think the data is easier to understand and use. The second is that it will be more memorable, and when something is memorable, people are more likely to use it and share it.
Take a look at the following two bar charts. Both contain exactly the same information. One of them appears in the IPBES report and the other was created by the Guardian’s graphics team to accompany their reporting of the IPBES report.
Figure 1: From the IPBES Summary Report (Figure 3A of the report).
Figure 2: From the Guardian.
At first glance they are very similar but there are four differences that make the Guardian version more pleasing to look at and easier to digest.
- The title tells you what insights you should be taking away from this table: A quarter of mammals and more than two-fifths of amphibians are threatened with extinction. Providing a verbal hint such as this has been found to increase people’s understanding of the data you are visualising.
- The key is clear, simple, and placed in a position where you will read it before you look at the table. Furthermore, the order of colours in the key follows the order of colours in the bars. The IPBES version of the key forces you to search for each colour and link it all together.
- The grey species icons have been deleted, removing unnecessary distraction and making the whole thing cleaner to look at.
- The number of assessed species has been removed from the right hand side of the bars, ensuring the reader focuses on the message highlighted in the title. Sometimes by giving people less information, they will remember more of what they see.
Overall, the Guardian’s version of the bar chart seems to be following the principles of Tufte’s data-ink ratio where only the ‘ink’ that’s needed to display the data is included. All other embellishment ‘ink’ is removed. The result allows the viewer to focus on the data.
Now compare the previous two data visualisations to the one below. Based on exactly the same extinction risk data, IUCN have used photos in this version to spark an emotional response.
Figure 3. From an IUCN Red List Brochure.
Just look at that little Tarsier’s big eyes. It’s pleading with you to save it.
A little emotion can go a long way when it comes to nudging people into action, especially when you choose images that will set off a chorus of ‘aww’s’.
Stripped right back, this IUCN data visualisation focuses only on the percentage of species within a group that are at risk of extinction. The red slice of the pie is literally wiping out the animal, plant or coral pictured beneath it. It packs a real emotional punch. You can see in one quick glance which species groups are most at risk and where help is needed. The simplicity of this data visualisation reduces any chance of misinterpretation — the plants, animals and corals of our world are dying and they need our help.
In comparison, the following line chart is open to misinterpretation.
Figure 4. Declines in species survival since 1980 (Red List Index). Figure 3C from the IPBES report.
Looking at that gently sloping line, you could be left with the impression that mammals aren’t doing so bad and we probably don’t need to worry about them. Corals on the other hand, jeeze…they aren’t doing so good. But this is interpretation is wrong. Mammals are at risk and many, many species are teetering on the knife edge of extinction. You just need to look at the accompanying graph of species extinctions to see that the rate of mammal extinctions has actually been rising rapidly since 1800.
Figure 5. Extinctions since 1500 for vertebrate groups. Figure 3B in the IPBES summary report.
In the IPBES report these two visuals are placed side-by-side, compounding the confusion of what the message actually is. Should we be worried about mammals or not? In comparison, Figure 3, the visualisation with photos, gives you a much clearer answer.
Choose just enough, not information overload.
Good data visualization helps us unpick complex issues, and the destruction and preservation of our planet is possibly the most complex issue we need to solve right now.
“Visuals are often easier and faster to process than numeric or semantic information. They make research easier to access, understand and use.”
As we saw in the previous examples, reducing the amount of information you put into a data visualisation makes the data easier to consume. On the other hand, the visuals in the following table do not make data easier and faster to process. The table contains lots of different pieces of information and the overall message is hard to pick out unless you read the caption that goes with it.
Figure 4. Figure 1 from the IPBES report.
This table is like white refined sugar — overly processed and void of nutritional value. It might look tasty but it’s going to leave you feeling unsatisfied. The main area of confusion is around what the arrows and circles mean — it’s not immediately obvious and you have to seek out the key which is located at the bottom of the table. In this example, the visuals have not made it faster and easier for us to understand the research and it would be easier for our brains if the table simply said ‘consistent’ or ‘variable’ to indicate the trend across regions.
The 50-year global trend symbols and arrangement is also confusing. Is it implying that the trends will slow down/increase in speed over time, or is the location across the x-axis an indication of when the worst of the trend will kick-in? When we look at a data visualisation, we shouldn’t be wondering how to interpret it. If we make our viewers think too hard, they’ll look away without taking the time to learn what we’re trying to tell them. This table would benefit from an application of Tufts data-ink ratio and Jordan Harold’s research on using semantics over symbols.
In an another example of information overload, Melanie and Dani both wondered why there’s a map in the figure below, and why it’s been placed in the middle of the table.
Figure 5. Figure 8 from the IPBES report.
Ignoring the question of why there’s even a map here for a moment, if the map intends to show where each sub-continent is located, then boundary lines are needed. The labels would also be easier to read if the font was smaller and in lowercase.
However, is this map even necessary? How many people need reminding where South America or North East Asia is? Does knowing the location of each sub-continent help us understand the projections of land use better, or is it merely a distraction? We say it’s not necessary, but if it is, it should be placed at the top or bottom of the figure so it doesn’t break up the display of data.
To improve the usability of this particular data visualisation, we’d play around with the bars: put all the orange bars together, flip them horizontally, and put the region labels horizontally next to the bars. This way we could compare the regional effects more easily and the eye doesn’t have to jump from orange, to purple, to white, over and over as it scans the page.
Choose good news over bad news.
The media loves a dramatic disaster story and headlines like “World on track to lose two-thirds of wild animals by 2020, major report warns,” and “We have 12 years to limit climate change catastrophe, warns UN” have had some success at stoking the flames of indignation.
When you consider the media’s lust for drama it’s easy to understand why the IPBES media release focused on “Nature’s dangerous decline”. But honestly, I’m tired of hearing the same old thing. Yes, we know there’s a problem. Yes, it’s us. Yes, it’s bad. And yes, there is something we can do about it. We sound like broken records and people are willfully ignoring us.
The problem with negative headlines is that they are overwhelming us and scaring us back into our safe little pillow forts where the monsters can’t touch us. Rather than dealing with the problem and focusing on what we can do, we’re allowing ourselves to be scared into denial, submission and inaction. The big oil companies must be delighted.
Policymakers are humans, just like the rest of us, and they want to hear positive news too. Let’s give them positive stories and show them what’s possible when people challenge the status quo. We have the data and research to support our ideas — we just need to present them in more accessible, useable and beautiful ways.
Choose anything but a PDF.
Creating an online report — or even better, an interactive version of it— could open up the latest IPBES report to more readers. Google indexes PDFs but they are rarely optimised for SEO. Add to that the fact that PDFs themselves make it hard to find the information you’re interested in. You have to scroll through to find the section you’re interested in or use the find function to search for keywords and click through each one till you find the relevant content.
With the right information architecture, content can be organised and structured in a way that makes it easy for people to find the knowledge they need and discover new insights they weren’t aware of. It could save people time and effort, reducing the barriers to access. We could even go one step further and give people the option to personalise their experience and focus on one particular region or area of interest, increasing the chances that they’ll do something with the data at their fingertips.
We also need to consider how people are accessing information. As mobile ownership rises, and the way people work changes, we need to consider the people who are accessing content on their phones or tablets. Is a PDF that you have to zoom in on and move around to read going to be the most accessible format?
Mind the (communication) gap.
Good data visualisation requires time, effort and money. But all too often dissemination is an after-thought, one that comes when a project is almost over and the budget already spent. Ideally, communication and feedback should happen throughout the life of a project. A steady stream of information can build momentum and grow into something that sticks better than a weighty tome thumped onto the internet in one go. It also provides an opportunity to adjust your approach, tweak your data visualisations, and ensure your research is being put in places where people will see it.
There are billions of pieces of content out there competing for our attention and we can’t expect people to come to us anymore. You have to take your research to the people and put it right under their noses. And if you want them to read it, you have to make sure it’s relevant, interesting and beautiful. Otherwise, they’ll scroll right on by.
It takes a team to create good dataviz.
And finally, as Melanie wrote in her last blog post, the creation of high-quality dataviz requires a team effort. To create something that’s easy to use, beautiful, and impactful, you need designers, writers, researchers, scientists and psychologists. Working as one global, multi-skilled team we can step into the smouldering world and save the amazing diversity of life that makes our planet so special.
So remember, when it comes to making data visualisations that capture attention and spark action, follow these four principles:
- Choose Beauty, not the Beast.
- Choose just enough, not information overload.
- Choose good news over bad.
- Choose anything but a PDF.
Melanie is a Human-Computer Interaction Researcher who aims to understand the psychology behind effective data visualisation and how we can make better, more impactful data visualisations. Melanie’s research is made possible through the Knowledge Transfer Partnerships scheme managed by Innovate UK, the UK’s innovation agency, and is conducted with the support of our partners in the Department of Psychology at Cambridge University.
Camellia is Vizzuality’s Lead Writer. She shines a spotlight on the digital technology that helps us better understand our world and how to protect it.
Dani is a Designer who’s always on the lookout for new techniques and technologies, seeking better ways to translate data into knowledge.
