In a guest post for OJB, Steve Carufel interviews Dutch data journalist Thomas de Beus about visualisation, storytelling — and useful new tools for data journalists.
Data journalism is, among other things, the art of resisting the temptation to show spectacular visualisations that fail to highlight the data behind a story.
Insights and relevant statistics can get lost in visual translation, so Thomas de Beus’ Colourful Facts is a great place to start thinking more about clarity and your audience — and less about spectacular graphic design (although you do not want to forego attractiveness entirely).
Thomas is a Dutch freelance data reporter who has recently joined de Volkskrant as a Data Researcher for the website’s graphics desk. He was also part of a startup named New Atoms which was granted funds by the Google News Initiative last year.
But it is with his blog — spotting and improving flawed data visualisations he comes across online — that de Beus has made a name for himself.
Because it is easy to be tempted by something visually impressive and beautiful, even for experienced data viz creators, here are 8 easy tips for improving your data viz in your daily practice.
I had to do it #DistractedBoyfriend pic.twitter.com/WkuISzzeP5
— Xaquín G.V. (@xocasgv) 27 août 2017
1. Don’t use fruits (or any irregular shape that relates to your subject)
Thomas advises keeping in mind that fruits are probably not a good option for visualising data:
Use avocado’s for guacamole not as diagram please @trouw #dataviz #infographics #ddj pic.twitter.com/xIzBFEQOdP
— Thomas de Beus (@TdeBeus) 11 mai 2017
While this may sound obvious to some people, it’s easy to give in to the temptation to entertain yourself. But you should not try to impress readers by showing your ability to make a fancy viz if this makes the data less clear by any margin.
For these reasons, Thomas de Beus insists on keeping the audience in mind by visualising the data in the clearest way possible. Here is another even more dubious example.
2. Speaking of food again, avoid pie charts if possible
After bars and columns, pie charts probably are among the most well-known types of visualisation. While they are easy to understand in some cases, they tend to poorly communicate the full picture:
Stumbled @AT5 upon horrific graphic from election results Amsterdam. Decided to improve it. Inspired by @alekswis @ftmedia. #dataviz #ddj pic.twitter.com/ZDt1seQomq
— Thomas de Beus (@TdeBeus) 16 mars 2017
They may be appropriate in showing simple proportions, but they are a poor choice for comparisons or trends.
Do not hesitate to rely on bars or columns since they often make the data much easier to assimilate, especially when differences in data points are slim.
3. If you go for a pie chart anyway, use it properly
The pie chart below and on the left is misleading because this type of visualisation is about proportionate shares, but the colours here designate mere categories. The actual data points are the multiple thin rays originating from the middle of the circle.
Stumbled upon the left chart or #dataviz by @OECD via @wef. I thought it could use a redesign and added the Netherlands. pic.twitter.com/hpuOPMvfOI
— Thomas de Beus (@TdeBeus) 31 mars 2017
The chart on the right may be less artsy, but it makes the data much easier to read and compare. This brings us to our next tip…
4. Put the data and its clarity above everything else…
Resist multiplying the insights within a single chart because complex visualisations can be a turn-off too.
While you probably have been working for a while on your visualised data, your audience is seeing it for the first time.
Explaining complex subjects that use data as efficiently as possible is what drove de Beus into data journalism.
Visualisation, again, is about making something clearer, more apparent to the reader. Our role is especially relevant when the whole data story is complicated in the first place:
“People are visually-oriented when it comes to consuming content. My core mission in journalism is to use data journalism as a tool to tighten the gap between reality and the people’s perception of reality. This is a huge problem right now and has always been.”
“I feel that, with data as a source, I can get very close to that reality. With visualisation, I try to help the reader in the best possible way.”
5. …But do make it beautiful so people will want to share it
On the other hand, being too minimalistic with your data viz will be counterproductive for the visibility of your story:
“In journalism, we need to keep our audience at the top of our mind. We serve a purpose, which is to explain a problem or story in the best way possible.
“I do however think there should be a healthy balance between the beauty and the effectiveness of a visualisation. For instance, data visuals are easily shareable on social media. If you want people to read your visual, it should be attractive too!”
In other words, prioritise the clarity of your data. Only then should you make it more attractive without altering its meaning, not the other way around.
6. Explain what the data means, not simply what it is
Unfortunately, visualising the data is only the first part of your upcoming story. As the data is exhaustingly scraped, visualised and made comprehensible, the process is still not finished. As storytellers, data reporters must now make sense of it by telling the readers what it all means.
“Storytelling is not a hot topic in data journalism for nothing. It is difficult to master. But I think the most important thing in telling a story with data is to empathise with your reader. Your story is not the data but what the data represents.”
“Data tells something about our lives and how we design it. Constantly ask yourself: why should people know what I’m telling them? And start by asking yourself why sharing a certain insight is important.”
7. For starters, small is beautiful
Thomas recommends to quickly and constantly publish little things, rather than spending too much of your learning time on bigger stories. In his experience, it was more draining than he expected.
Leave the lengthy reports for later. Data fluidity will come faster. Plus, updating your website more frequently with quality entries will help your search engine ranking too!
8. Iterate, submit to critical review, don’t be alone
Iteration and testing is an integral part of learning the creation of high quality, insightful visualisations.
“Storytelling is a skill you should practise and practise. Let someone read your raw ideas, sketches or paragraphs and iterate.”
Scraping, coding, analyzing, newsgathering and storytelling may seem like a lot. Remember that, in many newsrooms, the data unit is rarely a one-man army. And since the tools and trends are continually changing, data reporters must expect to regularly (re)learn:
“I agree with the fact that nowadays a data journalist needs to have a lot of skills (that’s why data journalism at big publications happens in teams), but hey, that makes it fun too! The thing I love most in life is learning new things!”
“Teamwork is my preference. You learn much faster.”
Steve Carufel is a student on the MA in Multiplatform and Mobile Journalism at Birmingham City University.
Reblogged this on Matthews' Blog.
Reblogged this on do not drop the ball and commented:
Data made visual – with a thought…
Pingback: Stunning Visualizations, Humanoid Robots, AI Politicians, and More: 8 Insightful Data Science News Articles from November 2017 - Kolabtree Blog
Pingback: Visualisation de données: des erreurs typiques à éviter – Steve Carufel
Pingback: Data visualisation: common mistakes to avoid – Steve Carufel