Accurat’s Perfectly Imperfect Approach to Data Visualization

Data isn’t perfect, despite appearances. Numbers, lines, and charts have a way of tricking us into believing they’re fact. The reality is far more messy.

“Data isn’t the infallible god that’s keeper of our truth,” says Giorgia Lupi, one of four partners at Accurat, an information design studio with offices in Milan and Brooklyn. Lupi’s words might ring as blasphemous to people who would rather believe that data is a key to unlocking insights about our chaotic, complex world.

But for Lupi and her three partners, Gabriele Rossi and Simone Quadri, and Paolo Labbozzetta data is more about depicting reality than it is about absolute certainty. “Different datasets have different ways to approximate reality,” says Rossi. “No dataset tells the whole and only truth.”

Accurat’s unorthodox approach to data has positioned the company one of the more intriguing information design studios in recent memory. Its work straddles the line between artistry and corporate art form, turning troves of data into beautiful, compelling, and often complex data visualizations. You might recognize Accurat’s designs from the interactive web app it made with Google. Or maybe from IBM’s data visualization language. Perhaps you saw it plastered against the wall at MoMA during the recent Items: Is Fashion Modern? show. Lupi is a star in the information design world thanks to her clever Dear Data project, which has accomplished the impressive feat of making data both interesting and relatable.

The company’s clients cut across scales, from Fortune 500 companies like IBM to smaller editorial partners like newspapers and magazines including National Geographic and Wired. The work ranges from straight data visualization illustrations to building technological tools that help companies collect and wade through complex mounds of information. “I like to say we’re a design agency crossed with a design consultancy,” Rossi says.

That hybridization approach to the studio’s work isn’t surprising given the diversity of the team’s background. The founders come from different fields, ranging from architecture to economics to interaction design. The larger staff is just as varied with a makeup that includes data scientists, analysts, UX designers, interface designers, and data visualization designers.

When Accurat first started out in 2010, it primarily approached data as another form of narrative storytelling. Its early work centered around an ongoing column it had in the Sunday supplement to Corriere Della Sera, one of Italy’s largest newspapers. Accurat’s then-small team worked with the newspaper editors to create narrative data features that delved into cultural issues like education and life expectancy, and the patterns hidden within the best movies of all time.

The stories were printed for a mass audience, yet there was nothing easy about them. The illustrations tended to be dense with information. There were colors, shapes, and intersecting lines, all meant to represent different facets of the dataset. Each came with a legend that helped readers wade through the complexity, which Accurat, on principle, doesn’t shy away from. “In general, we tend to create very rich visuals that sometimes need a little time to be read and understood,” Lupi says. “We’re trying to create an engaged form of reading.”

As Accurat grew, Lupi, who runs Accurat’s design team, began rethinking how the company should approach data. Plenty of companies were able to make beautiful infographics-what they lacked was a perspective on what it all meant. This was around the time Lupi was making Dear Data, the project in which she and the designer Stefanie Posavec turned the minutia of their daily lives into striking data visualizations. Lupi realized that she and her team could bring this idea of “soft data” into more of Accurat’s client work as a way to help companies glean more nuanced truths from data sets. “Working on a more artistic exploration helped me focus on what data is: a medium to describe our reality,” she says.

Last year, Lupi published a manifesto of sorts, outlining her and Accurat’s more “humanistic” approach to data. In it, she describes a less technical approach to data visualization; one where qualitative information (human sentiment, for example) is just as important as the quantitative (i.e. earnings report numbers). If done correctly, the two can live alongside each other-the fuzziness of one adding nuance to the hardness of the other.

“People aren’t used to seeing data as something imperfect,” Rossi admits. But there’s a crucial distinction to make between inaccuracy and the imperfection that’s bound to show up in any human-gathered dataset. In many cases, accurately depicting a dataset means including its flaws-its biases, its approximations, its uncertainty. For Accurat, those flaws are where the data becomes interesting. It’s an opportunity to make data feel more human, beautiful, and true.