For graduated astronomer Nadieh Bremer, creating data visualizations in celebration of the Hubble Space Telescope’s 30th anniversary in orbit was a dream come true. Located in a small town near Amsterdam, the data visualization designer was approached by Physics Today magazine with a dataset of all the science observations Hubble had done – more than 550,000 in total. Unsure of what exactly they wanted to visualise at that point, the project eventually became one of the biggest ones Nadieh had ever worked on and her proudest achievement in data visualization yet.
“It was almost like a personal project for which I was being paid,” Nadieh recalls. “Hubble’s amazing photos were one of the main reasons that I fell in love with astronomy in the first place, so I was beyond excited to work on this project. I spent a lot of time on it, way more than the budget allowed, because I was just having so much fun with the data. I really wanted to make it as special as I possibly could.”
The size of the dataset made it clear that the angles to visualise could be endless, so Nadieh cleaned the data and performed an extensive analysis before pitching four story ideas to the client. They settled on two static visualizations – a main one that would plot all of Hubble’s observations in the sky and highlight some of its most famous photos, and a companion piece that would reveal the diversity of the objects that Hubble has explored.
Physics Today was so enthusiastic about the main visual, which was originally only meant to appear on its website, that it was adapted and also published as a gate-fold poster in the magazine. (For detailed explanations on how Nadieh created both visualizations, see her extensive blog posts on the sky map with all of the observations and the different astronomy targets. Prints of the sky map are also available in her online shop.)
“I love having this treasure chest of possible stories and insights that I can tease out through data analysis and client conversations. I really pick apart the data and look at it from different sides.”
Uncovering the stories hidden in the data
Over the course of the last three years, Nadieh has carved out her own niche as an award-winning freelance designer of beautiful yet also effective and user-friendly data visualizations. Created with a combination of first class technical skills and a very creative mindset, the end result is utterly unique and memorable every single time.
One of Nadieh’s favourite parts of the work is exploring a dataset and discovering the stories within. Usually, the client has an idea of what they’d like Nadieh to focus on but – as was the case with the Hubble project – sometimes they invite her to really dive in and find interesting angles herself.
“I love having this treasure chest of possible stories and insights that I can tease out through data analysis and client conversations,” Nadieh explains. “I talk to the client to figure out the goal – what they’d like people to do, learn or get a feeling for when they see the visualization. Then I really pick apart the data and look at it from different sides.”
Nadieh uses statistics language R to prepare and analyse the data, which helps her build up a mental model of what it’s about, find the most interesting aspects that lend themselves to a visualization, and get a feeling for how it could be restructured.
Once Nadieh has a good understanding of the data and the goal has been decided on, she starts thinking about the basic visual form, a critical step in creating an effective visualization. She gets inspiration from Pinterest boards that she curates, which contain other data visualization work, colour palettes, visuals from space and natural history, spirograph art, and much more. Before she starts designing anything, she creates a client mood board and adds everything that could be relevant for the project.
She also takes the dataset itself into account for the design. For example, the visualizations for a series of data art posters that Nadieh recently designed for Sony Music were inspired by the aesthetic of the traditional gold record. The pieces use data from Spotify and YouTube streams, chart entries and the audio characteristics of the songs itself to present each song’s data-based “fingerprint”.
Behind the scenes of Nadieh’s design process
After preparing the data, Nadieh then enters the design phase. She draws rough sketches – either with pen and paper or the Tayasui Sketches app on the iPad – to explore possible visual directions. Circles feature heavily in Nadieh’s visualizations, as do vibrant colours.
It’s this creative part of figuring out the visuals and seeing them come to life in a way that fits the data, goal and topic that Nadieh enjoys the most. She spends hours experimenting and tweaking the design to make the final visualization both aesthetically pleasing and effective.
“The reason I only create very rough outlines at the beginning is that data visualization stands and falls with the data itself,” Nadieh points out. “You can have a great idea but when you plug in the data it sometimes doesn’t work at all, because of outliers in the data or other quirks that you hadn’t considered during the design process. When you work with big datasets, you can only guess at the start.”
Nadieh implements her ideas in Visual Studio and codes with D3.js – the tool that’s become the standard for building data visualizations online. At the beginning she says the visual is still very ugly but it helps her see if it has potential. Once she’s confident that the idea works, she’ll go through lots of iterations.
“I try and add as much context and nice-to-haves as possible,” she explains. “I play with the colours, opacity, shapes and other effects to make the visualization more interesting, and the more invested readers can find multiple stories. Six months later, if I did it again, the same visualization would probably look very different. What drives me is how pleased I am with the end result and if it’s effective.”
Depending on the amount of data to show or the complexity of the interactions, Nadieh will also use SVG, HTML5 Canvas, three.js or GSAP. She keeps things simple and tests in Firefox and Chrome straight away.
“I don’t want to create dependencies for the visuals I create,” Nadieh reveals. “Making my interactions as low level as possible means the client can view it on their system and hopefully make it fit whatever tool they want. There’s no React or Vue underneath it. It’s quite minimalistic.”
This is where things become more challenging, because the visualization needs to work across various browsers and devices with different screen sizes. Nadieh finds that working on static visuals is often more enjoyable, as she can focus on what interests her most, as opposed to worrying about the technology.
“Having a really tiny space to tell a story puts a big constraint on the creativity,” Nadieh sighs. “Often the interactions on mobile have to look different to the ones on desktop, and there’s just a lot of extra stuff that needs to happen to make it work. Also, there are performance considerations – load times become an issue and animations don’t look staggered anymore – and so many browser bugs. It’s more of a technical than a creative challenge.”
From studying astronomy to signing up Google as a first client
Nadieh has been fascinated with astronomy since childhood and went on to study it for five years at the University of Leiden. She hated writing papers, however, and after graduation joined Deloitte Consulting's new analytics department.
“I was a data scientist before that role was really known in the Netherlands,” Nadieh remembers. “I cleaned, assessed and analysed data from external companies and investigated customer segmentations. For client presentations I created charts and discovered that I loved the visualization of the analysis and insights. As things became more complex, I learned how to make my visualizations more professional, style them and add interactivity with D3.js.”
Nadieh, however, started losing her enthusiasm for data analysis and wasn’t looking forward to creating yet another predictive model anymore. Seeing a speaker, Mike Freeman, call himself a ‘data visualization specialist’ on the first slide of his presentation at a data science conference was a pivotal moment.
“It hit me like lightning,” Nadieh laughs. “You can specialise in data viz?! From that moment I knew it was where my passion lies and started to dedicate all my spare time to learning more about data visualization. I love the synergy between the math side and the creative side.”
In 2016, Nadieh also began a year-long collaborative side project with fellow data visualization designer Shirley Wu called Data Sketches. Every month Nadieh and Shirley – based halfway across the world in San Francisco – would choose a topic (from movies to myths and legends), create an elaborate data visualization each, and document the entire creative process in meticulous detail.
The project, which is currently being turned into a book to be published in early this year, put their names on the map beyond just the data viz community. It won an Information is Beautiful award the following year (Nadieh also took home ‘Outstanding Individual’), and led to Nadieh taking the plunge to go freelance. Her first client was Google.
“Alberto Cairo, a big name in the data viz community, was art directing a series for Google News Lab,” Nadieh explains. “He reached out and invited us to create visualizations from Google’s huge search term data. I didn’t want to say no to this amazing opportunity, but it meant I had to go freelance. I had been playing with the idea for a while, and I thought having Google as my first client would be a very good way to start!”
For her contribution, Beautiful in English, Nadieh analysed the most popular words translated into English through Google Translate (while Shirley dug through a decade of travel searches from nearly 40 countries to explore the top searched cultural locations).
Nadieh has freelanced under the name Visual Cinnamon ever since. It was one of the best professional steps she’s taken and hasn’t regretted it once. Her impressive client base ranges from small startups to household names such as UNESCO, the New York Times, and Greenpeace. The topics, meanwhile, are as diverse as the huge number of digital trackers that you gather when you browse the web, the sales of highly hazardous pesticides, and the questions that people ask on Google to understand their cats and dogs better. Because of her background in astronomy, clients also approach her more and more for space-related projects.
Handling large, complex datasets
Nadieh usually works on three to four visualizations at the same time, so she can switch between them and not get too overloaded with one project. Each visual is custom-made for marketing (both internal and external), press releases, and articles in print or on the web.
Lately, the datasets Nadieh is working with have been getting bigger and bigger (from around 100k-600k up to 60 million data points). She enjoys the creative freedom of figuring out how to visualise the diverse data, discover additional stories besides the main insight and show the context.
“When datasets are that big, you need to handle them on a case by case basis,” Nadieh explains. “Sometimes you can actually show 600,000 data points. I made a visualization for [Earth imaging company] Planet that showcases their fleet of small satellites going round the Earth and the 600,000 images they capture every day. But now I’m working on a project with 60 million data points for the Swiss electricity transmission network. It’s the data for every hour in 2019, and I have to figure out what to prioritise and aggregate, because there aren’t 60 million pixels on a typical screen!”
Another current project focuses on visualising brain bleeds, using data from thousands of MRI scans to better understand how bleeds might affect different demographic groups. Then there’s a visualization for a report on the impact of COVID-19 on retail, and of course another space-related project – about all satellites orbiting the earth and the possibility of them interfering with each other – is also in the mix.
Her innovative projects have won Nadieh numerous awards and made her a sought-after conference speaker. It’s easy to see why. She has a knack for revealing the stories hidden within large and complex datasets in a way that captivates, engages and enlightens audiences around the world. Whatever the subject, Nadieh always succeeds in making her visualizations both beautiful and effective. But it’s also her incredible passion for her work that sets her apart. In everything she does Nadieh demonstrates that designing data visualizations can be so much more than just creating standard charts and graphs.