Nadieh Bremer is an award-winning data visualization designer and artist, from a small town near Amsterdam. Nadieh’s projects demonstrate the beauty, fun, and excitement of working in dataviz. A few weeks ago, Nadieh shared with us a behind-the-scenes of her latest work as well as some of her best practices and inspirations. NB is Nadieh Bremer, NR is Noëlle Rakotondravony (DVS), JF is Jason Forrest (DVS)
JF: Why don’t we get started with a quick introduction for our readers.
NB: I’m Nadieh; I graduated in Astronomy, then became a data scientist. When I was doing analytics and machine learning, I figured out that I really like data visualization. So I jumped in, trying to learn as much about dataviz as possible, reading books, learning D3.js, and best practices by making lots of personal projects.
I started out as a data scientist at Deloitte, but after about 4 years I switched to Adyen, where I was mostly designing and creating d3-based dashboards. But that wasn’t quite the dataviz I like, so I became a freelancer to focus a lot more on the creative side of dataviz.
NR: As a dataviz freelancer, what has been the coolest thing you experienced? What was the least enjoyable?
NB: The coolest thing, in general, is the diversity of projects. I usually have two to three clients at the same time ranging from airlines to startups to pharmaceutical, investments, UNESCO, and Google. I like being able to flip between different datasets which have their own insights that I try to tease out and visualize. That gets my enthusiasm really high! The thing I least enjoy is setting up contracts which can take several hours before it’s done properly. I wish that part could be skipped!
NR: Which of your latest works are you the proudest of?
NB: Fromthe latest ones I really like what I did for Google, “Why do Cats and Dogs…?”, which investigated the most popular questions that people search for in order to better understand the behavior of their pets. It was a lot of fun to do and I got a lot of creative freedom to make it.
JF: When they gave you the premise, was it just something like ‘do something with our search data’ or was it specifically about pets?
NB: It’s basically: “do something that uses Google data.” This doesn’t even have to do with Google’s search data. However, they do like it if you use and incorporate Google trends in one way or another. You’re free to come with ideas and create the designs that seem most interesting.
JF: At the EYEO Festival, you talked about how you made the clusters as different little universes by starting word. There are so many types of visualizations in that one project, how did you come up with the structure and story?
NB: It was supposed to be much smaller. I wanted to create a sentence tree, it’s a flow chart from the center out to create a sentence by following lines through words. For example, why does my cat — like — to bite — my hand? If you follow different words you might see a different question. This can show how popular these questions were.
The dataset was so big with too much information that this idea didn’t work anymore. So I ended up with these sort of universes: one for each of the branches of the main tree. But that visual was still too big and I had to guide people into it to show all the data. That’s when the separate bubble idea came about: to make it easier for people to grasp the concept of these sentences that they’re looking at.
The next visual looks at the difference between the most asked questions that include the word “my” (versus not including “my”) like my cat, my dog. That was already in from the start because I also wanted to look into other aspects of the dataset. Then I wanted to end with Cats vs Dogs as a final visual. Some people have cats, some have dogs, and in movies, it’s always cats vs dogs, so that seemed like an interesting way to wrap everything up.
NR: After preparing your data, where does your inspiration for the visuals and colors come from? Do you follow some specific steps?
NB: First, I need to know what the visuals should convey to the audience and what should they learn. I also need to understand the data, and I always ask from my client a sample or the complete dataset so I can get a feeling for what variables and values are in there.
Then I start making very abstract designs to illustrate the topic that I always keep in the back of my mind. What I draw here could be anything. Sometimes I color the raw sketch to make it more visually appealing to a client. I also navigate my Pinterest boards during the design phase; I usually create a “client board” with things that I think would work for this particular case, goal, data, and client, that seem interesting — even fascinating! Apart from that, it’s totally free form: I start experimenting and then I get into several iterations until the point where I feel ok.
As for the Cats and Dogs, I wanted something different than the perfect circles that I’m usually using. I hired Juliana Chen, an animator, specifically because I liked her cute style, how she also sometimes uses rougher brushes to draw with. So I did a few different things before finally ending up on the sketchy style. Sometimes the whole process can involve a lot of googling, looking around and having an open mind, knowing when something you see seems to go in the right direction.
NR: Did it ever happened to you that after the release of a project people did not get it or could not feel connected to the visuals that you built?
NB: Yes and No. For client projects, I never had people saying that they had issues because our iterations provide multiple opportunities to get feedback and I respond to that. When we get to the final result, they have already shown it to other people in the company and tested it themselves.
With personal projects, I usually have one or two people who give me feedback. I don’t think I have had a project that completely caused a facepalm but there were definitely things where people were like ‘I don’t get this’, ‘this visual form is not as effective’ or ‘I would do things differently’. But just showing it to one or two people can already get out so much of the worst mistakes that it’s extremely valuable to have that.
I also hear people say that they don’t understand my visualizations. That’s because I make more complex visuals where you need to give a little bit of effort to understand what’s shown, in order to gain a lot of information eventually. Some people just don’t want to do that.
JF: What is it about these complex data stories that you find so fascinating? I’m curious why you focus on presenting them in a way that’s counter to the ‘data:ink ratio’?
NB: I like it when I can give people the opportunity to explore the main story through different angles. For example, I made a visual about the Olympics and its gold medal winners. That was five thousand different medals that I tried to display in a way that made it easier to see trends. I like the fact that everybody can look at their own favorite parts of the main story and see insights like how some sports completely flip from always literally European winners to Asian winners and other smaller things that happened.
So while showing the complexity of the data, people can get more context and a complete idea of what’s behind the average numbers. They can look at things that they find interesting instead of things that the designer has deemed as interesting.
The second part why I like diverse datasets is because it gives me a lot of freedom to be creative, to play around trying different things and make visually unique graphs. When I have a hundred data points with different sizes, I can make something more visually appealing than if I have fewer data points and sizes. Larger datasets definitely challenge me more on a creative level!
JF: So you said that it helps you to make a more unique solution. Is that the focus for you each time to try to do something — that it is unique?
NB:Well actually what I want is to make my visuals custom, appealing or attractive; and by doing that, they eventually become unique. Unique is just a word but it’s not the goal. By having diverse data, there are more ways to show the uniqueness of the data and the stories in it. Those are goals that are heavily drilling in my mind and the way how I approach it. Sorry, that’s not a really good answer.
JF: It’s really fascinating! How to make something dynamic, appealing, and compelling but easily understandable is something that I’m preoccupied with as well. I think the problem that we often have in data visualization is that our audience wants a dazzling visual that is beautiful but it disappoints if they can’t explore it.
NR: One next question is which tools do you use to make your visualizations?
NB:These are my basic tools: R, D3 and the web, and Illustrator — and of course, pen and paper! First, I do lots of ugly charts in R to get a feeling of the data: ‘what is the minimum, the maximum, the distribution of the variables, their values’. Then I restructure the data in R to make it into the shape that I need to create my design. For now, I’ve always been given the freedom to restructure the data a little bit to make it fit the visualization.
After I have the dataset in the right shape from R, I will move to the web & D3. To me, D3 itself is more the building blocks to make the visualization possible. Then I create it using SVGs or if it’s a lot more data I use Canvas, and if it’s like a crazy amount of data I might go into WebGL. For static things that end up in print, I will export the (SVG based) data visualization into Illustrator and add things that are harder to code like adding annotations, legends, etc.
JF: So we talked a bit about color, and that’s one aspect that sets your work apart. Can you talk about the vibrant palettes that you often pick? How do you assign colors with variables? What role does color have in your work?
NB: When I start, I have this default rainbow palette that I’ve been using for years now, just to make the screen more fun to look at. Then while I’m setting up the basic shapes, I might think of different color palettes. For example with companies, I try to make a palette that’s based on the colors that they have in their own house styles. Typically the clients don’t mind if I use a little bit more variation as long as the main colors shine through. But I do like bright colors because I personally find them beautiful.
I often end up with different palettes than the rainbow I started with. Few projects stick with my default rainbow palette because I couldn’t come up with anything new. But in the meantime, I would probably have tried like twenty different color palettes before settling on the one I find most suitable for a particular project. It’s always hard, sometimes you only need five colors and sometimes more.
The Cats and Dogs project was a fun one because there were no variables in my data that I could assign to color. But I didn’t want to make everything the same color, so I assigned color randomly from an underlying palette. Which meant that I could finally, for the first time, create a color palette where the color did not have to be easily distinguishable. There are two slightly different pinks and two different blues and greens — that was nice to do as a project.
NR: When building your visualization, do you have role models you feel looking at?
NB: Well, definitely Giorgia Lupi. I think what she is doing in terms of her career and her visual output is amazing. She has such a nice way to combine the design side with data visualization. These days as I’m venturing a little bit more into data art in general than before. I’ve been looking a lot at people like Matt DesLauriers and Manolodie.
NR: And our last question: as a member of the Data Visualization Society, what do you most look forward to?
NB: I find it fascinating! The combination of having the write-ups from different topics which digest some of the more important things is a great aspect I think. In the Slack, I am more of a spectator, I like peeking in the #showcase channel and see what people are doing and sometimes I’m included in the discussions which I find interesting. I wish I had more time to go through the Slack though. I also really enjoy the newsletters that summarize the most important things.
JF: I think you are exactly a good member! One of the things that the group is trying to find is exactly how to serve its membership. At this point, many members are not active each day but consuming the information and the discourse in a way that we still don’t fully understand — and that’s totally fine.
Alright, Nadieh, thank you so much. We really appreciate your time, love your work!
NB: Thank you for asking me for the interview, it’s very cool to be featured 🙂
Noëlle Rakotondravony is a Ph.D student in Computer Science at WPI. Her research focus is at the intersection of culture x HCI, data visualization, and natural language. Prior to WPI, she was a cybersecurity researcher at the University of Passau in Germany, which she joined upon graduation from the Telecommunications Engineering School in Tunis, Tunisia. Noëlle is also a polyglot, an editor for Nightingale, and a proud member of IkalaSTEM and its editorial team. Connect with Noëlle on Twitter or LinkedIn.
Jason Forrest is a data visualization designer and writer living in New York City. He is the director of the Data Visualization Lab for McKinsey and Company. In addition to being on the board of directors of the Data Visualization Society, he is also the editor-in-chief of Nightingale: The Journal of the Data Visualization Society. He writes about the intersection of culture and information design and is currently working on a book about pictorial statistics.
