Interview: Carnegie Mellon Professor Ari Lightman On How Students Are Empowered By Learning To Use Brandwatch Consumer Research
By Kara FinnertyJun 10
Published February 22nd 2017
A data visualization, when done well, can be an incredibly powerful way to communicate information.
We talked to master informatician and author of A Practical Guide to Designing with Data, Brian Suda to learn about his take on designing data visualizations that engage the reader and tell powerful stories.
To break the ice, he told us a bit about the origins of his interest in storytelling with data.
Suda did an undergraduate degree in software systems and computer science, then studied for his Master’s in Edinburgh, Scotland. He then went on to do another computer science degree.
“A lot of the data visualizations and the design come from more rigorous either mathematic or programming kind of backgrounds. But a lot of the work that I’ve been doing throughout the years has been on not-so-strict numeric information and numbers,” he explains.
In his book, Brian says:
“Much like you create a narrative in words, illustrations of data need to tell a story.
Keen to learn about the basic principles of telling a story with data successfully, we asked to hear more about it.
“There are lots of different story-telling techniques, and this is something that you don’t necessarily learn as a designer or as a programmer. I think bringing some key learnings from other disciplines like writing, both fiction and non-fiction, into data visualizations is definitely important.”
Some of the key principles he mentions center around answering a few simple questions.
Once you have those answers, you can work backward and figure out what strategy to adopt, which tools to use and what the desired outcomes are.
Suda thinks the best people in data visualization right now are journalists.
“First, they’ve got a lot of talent in writing already and they know how to tell stories, and they’re collecting this sort of data all the time. And then finally, they also need to push it out to their customers in an interesting way. So I think organizations like The Guardian, The New York Times, a lot of US newspapers at the moment are doing fabulous jobs at visual storytelling.”
When discussing common mistakes Brian has noticed in visual storytelling, he explains: “Several years ago, Mozilla had a competition. People opted in to have all of their activity in the browser tracked. Then it was anonymized, put into big Excel sheets or CSV and anyone could download and visualize this information.
Some of the results took every single column and every single row in that database, in that Excel sheet, and tried to make a visualization of it. And while that’s probably really good for some sort of business intelligence tool, there was no story.”
The challenge was figuring out what the data visualization was trying to tell the viewer and what was its purpose.
When working with big data sets, you have the permission not to visualize everything, but pick the most interesting information and focus on it:
“It’s okay that you threw away 99% of all the other data because you found this really tiny, interesting nugget that you wanted to tell a story about.”
In Suda’s eyes, the first time you see a data visualization it needs to tell you something at a glance and catch your attention.
“Maybe a boring black and white bar chart is perfectly legitimate in conveying the information, but if it doesn’t quickly grab my attention or make it memorable, then maybe it hasn’t served its purpose.”
Secondly, he emphasizes the importance of annotations and simply explaining what the visualization is: “I think a lot of people just fail to add a title. So you might see in a newspaper a beautiful line graph and it catches your attention, but then you wonder what is this? Is this GDP? Is this unemployment? Is this sign-ups?”.
Thirdly, Suda appreciates being told upfront what the data shows. So, rather than going for “This is a chart of the types of ice-cream that children like”, he’d prefer “Chocolate was the favorite ice-cream flavor in our survey. Here are the results”.
He concludes, “even if I don’t have a whole lot of time to digest the data, you’ve told me the key message in that headline itself”.
Ensuring that relevant information such as company KPIs is highly visible and accessible to everyone is highly beneficial and gives everyone an increased sense of responsibility.
If this information is kept hidden, it makes it very difficult to know if you are making a difference in the company. Transparency makes it easier for people to monitor the success of the organization.
One of the other great aspects is that by giving employees access to data they start to observe and make interesting correlations.
For instance, one person might say, Wow, every time we get a spike in tweets, it looks like a day later we get loads of new sign-ups. And then someone says, Oh, yeah, I wonder if those two things are correlated? And then someone in the BI team or data team will analyze the numbers and say there’s a 24-hour lag between people seeing it and signing up.
Without tracking that information on the big screen, no one would have asked those key questions. After finding out there is a correlation between tweets and sign-ups, they can use that information to inform important business decisions.
In this case, “it can justify hiring a dedicated social media manager because we know the more people talk about us, the more sign-ups we get. That’s something that never would have come out if an employee hadn’t looked at that board and asked the question.”