I was once auto-rejected from a job I had written for myself. After several interviews, my soon-to-be manager asked me to delineate the responsibilities I was eager to assume in the department’s new dataviz role. I responded with a detailed description that was subsequently posted on the organization’s website. As I went through the formality of applying, I encountered a question that human resources had added: Was my undergraduate degree in statistics, math, or a related field? No. (I majored in poetry.) I clicked submit, and was immediately rejected by an automated email. The gist was, Thanks for your interest, but you’re not qualified.
Ironically, my humanities background was exactly the reason the director had asked me to join her department. The difference in my educational training meant that, as an analyst, I complimented their existing STEM expertise instead of duplicating it. As a team, we could cover more ground. Many years and many projects later, my best contributions as a data visualization consultant are still consistently derived from my roots in the humanities. Thought leaders like Gioriga Lupi have powerfully advocated for the integration of data humanism to propel our work beyond superficial infographics and towards deeper, more meaningful visualizations. But despite the dataviz community’s celebration of such manifestos, there’s a lag in application. This chasm between theory and practice seems tied up in the composition of our teams. When I look around the table, I’m typically the only analyst without a STEM degree.
The data tells us that, left unchecked, people tend to hire others similar to themselves. The data also says, this isn’t just discriminatory. It’s also a business mistake, because more diverse teams produce better work. We should strive to construct teams that span, among other spectrums, the gamut of age, gender, race, and ethnicity. I propose one more diversity component we should consider as we build our teams: educational priming. If homogeneity hinders our best work, then the vast majority of data visualization teams and our projects are held back by academic uniformity.
If homogeneity hinders our best work, then the vast majority of data visualization teams and our projects are held back by academic uniformity.
To realize our full potential, dataviz teams (and for that matter, all teams in technology) need humanists. By humanist, I mean the folks who specialize in things like literature, philosophy, ethics, classics, history, culture, language, linguistics, the arts, and, my own first love, writing. (Throughout this piece, I use the term “humanist” to mean individuals with expertise in the humanities disciplines, analogous in functionality to terms like “scientist” or “social scientist.” There is no intended religious affiliation with secular humanism.) Humanities professionals have complementary skill sets that shore up some of the most common weaknesses in data visualization. As your dataviz team embarks on your next project, actively seek out a humanist. Here are the top seven reasons you’ll be glad you did.
1) Humanists create art
Sculptures, paintings, screenplays, poems: the world’s artistic creations are brought into being by humanists. We’ve trained to tap into that innate, deeply human gravitation towards beauty. And I do mean we’ve trained: practiced, drafted, rehearsed, again and again. Just like technical skills, aesthetic expertise does not materialize overnight. It takes years of effort to hone such a craft, and humanists have put in that work.
So often a dataviz team’s leading analyst is the member who possesses both technical and aesthetic talent. Folks say: “She’s a unicorn,” “God gave to her with both hands,” and then write it off as an unrepeatable anomaly. That’s a mistake.
Data visualization sits at the intersection of logic and art. Success hinges on visual impact as much as technical competency. What’s surprising, then, is that our dataviz teams don’t explicitly seek out folks who have developed these aesthetic skills. Designers, artists, and user experience experts should be pursued with the same dedication as statistical analysts, database specialists, and data scientists. So often a dataviz team’s leading analyst is the member who possesses both technical and aesthetic talent. Folks say: “She’s a unicorn,” “God gave to her with both hands,” and then write it off as an unrepeatable anomaly. That’s a mistake. Teams can and should consistently create that sort of synergy. They just need to intentionally couple their technical experts with colleagues specifically trained in the visual arts.
2) Humanists curate with confidence
Humanists excel at sequencing. We order essays into books, paintings into exhibits, movements into symphonies. We commit to a path. We arrange with a steady eye on the accumulation of meaning, calibrating how this moment converses with the preceding and subsequent piece. And if a segment doesn’t productively contribute to that arc, we set it aside. (This is good. You wouldn’t want to sit through a movie lugging around every scene floated at the writer’s table.)
This process of curation spooks the average analyst. Exclusions of any kind get conflated with dishonesty, and in the name of transparency, every chart, category, and filter ends up crammed onto the dashboard.
This process of curation spooks the average analyst. Exclusions of any kind get conflated with dishonesty, and in the name of transparency, every chart, category, and filter ends up crammed onto the dashboard. The resulting chaos repels the viewer like a museum would if their entire collection were shoved into a single room. Our audience should not be responsible for picking through all potential insights. That’s our job as analysts. As Cole Nussbaumer Knaflic put it: the audience doesn’t want to open all of the oysters; they want the analysts to do that, and then just show them the pearls. Hiding a pearl is dishonest. Setting aside an empty oyster is removing a distraction, clearing a productive path optimized for accurate interpretation.
3) Humanists write well
In the humanities, the vast majority of analytic output takes the form of writing. Typically there aren’t lab results, charts, or code—just ideas captured in essays, journals, and books. As students, when we turned in a crummy essay, we weren’t told to go rewrite it. We were told to go rethink it. Sort out our thoughts. Get organized. Clarify the thesis. From there, the words would fall into place.
The visual medium can become a crutch. If folks are brave enough to ask, “But what’s the point?”, we’re tempted to simply gesture towards the visualization and say: “It’s all there. You figure it out.”
This push towards clear writing forces humanists to fully articulate, for ourselves and our audiences, why our work matters. Many visualizations never have to cross that basic threshold. The absence of any written explanation means analysts often shortcut the difficult work of refining our thesis and building our argument. The visual medium can become a crutch. If folks are brave enough to ask, “But what’s the point?”, we’re tempted to simply gesture towards the visualization and say: “It’s all there. You figure it out.” A humanist on the team wouldn’t let that fly, and the end result will be better work.
4) Humanists question how we know
Throughout history, each era has leaned into the latest way of knowing. Today, science and technology are understandably at the forefront. But as we barrel along making such rapid progress, we sometimes forget that our latest and greatest ways of knowing are often riddled with errors or half-truths. As we talk about data, the confidence of our language too easily veers off into dangerous waters. Things like: Just show me the data. Numbers don’t lie. Hard data. Raw data. The data speaks for itself.
Humans seem to desperately crave knowledge they didn’t have a hand in making, some objective truth they couldn’t have accidentally muddied up. But this isn’t it. Data is a human creation, with all our brilliance and our blindsides baked in.
Humans seem to desperately crave knowledge they didn’t have a hand in making, some objective truth they couldn’t have accidentally muddied up. But this isn’t it. Data is a human creation, with all our brilliance and our blindsides baked in. As Lisa Gitelman succinctly put it: “Raw data is an oxymoron.” Data is just as fallible as any other form of human knowledge. This makes humanists especially valuable teammates: fundamentally, we study human creations. These folks are uniquely adept at spotting missteps and pointing out what we don’t know. As dataviz teams disseminate our era’s latest form of knowledge, we need skeptics at the table ready to question a flashy dashboard’s analytic integrity before it heads out the door. Humanists are great skeptics.
5) Humanists thrive on ambiguity
Ask a humanist, “Hey, does this belong in category A or category B?”, and they’ll inevitably frustrate you with, “It’s kinda both,” or, “It’s really neither,” or the ever-popular, “Well that depends on how you look at it.” For a data analyst, ambiguity is often the obnoxious hurdle standing between you and a productive dataset. For a humanist, ambiguity signals fertile ground for analysis. Simply put, humanists love grey zones. Continuous spectrums. Caveats, exceptions, paradoxes. That’s what makes the world so darn interesting. Fundamentally, that’s what humanists study.
For a data analyst, ambiguity is often the obnoxious hurdle standing between you and a productive dataset. For a humanist, ambiguity signals fertile ground for analysis.
And to be clear, we’ve got loads of STEM colleagues who value ambiguity just as much. But typically their training strives towards exactness, not variety. They produce knowledge through isolation. To the best of their ability, they try to freeze or control the context surrounding their research question, in order to determine whether A causes B, B causes A, or C causes A and B. Conversely, humanists revel in the proliferation of possible interpretations. To us, the world isn’t just multivariate. It’s like cotton candy: innumerable intricacies swirled into an inextricable knot from which there is no single thread to productively tease out. Context is everything. People who see the world like that prioritize different things. In data conversations, they push for more comprehensive taxonomies, more nuanced analysis, more pliable datasets that, yes, aggregate meaningfully, but also flex with reality.
6) Humanists remember we’re studying real people
When we study people, a dataset enables us to zoom up and out of our human granularity. We leave our singular occupancy of one perspective, one place, one time, and can suddenly see millions of lives at once, not as a crowded blur but as a meaningful whole. This is extraordinary. And of course, this superpower has a flipside. Distance is one of the easiest routes to dehumanization. Once folks are up in the spreadsheet stratosphere, amnesia is everywhere. It’s so easy to forget that each tiny box represents real individual human beings. And in being part of our dataset, seldom by choice, these people are vulnerable to our analysis, conclusions, and distribution of their lives.
Distance is one of the easiest routes to dehumanization. Once folks are up in the spreadsheet stratosphere, amnesia is everywhere.
Humanists, on the other hand, rarely mingle outside our mortal granularity. When our work is about people, it’s typically tethered to distinct, recognizable individuals. We know their names. We know their stories. Our best work is rooted in empathy, recognition, depth, even if the individuals we study are fictional or dead. Our goal is to get inside people’s heads, not to tally them up. Data visualization teams need more of this empathetic anchoring. An enormous amount of our visual analysis is conducted by folks who are not in the data themselves and who lack any personal connections with those who are. There’s often a significant power differential between the researcher and the researched, the visualizer and the visualized. And too often that’s paralleled by an intended audience who is similarly absent from the content. We can do better. Adding a humanist’s perspective is an excellent place to start.
7) Humanists know our analysis is of our time
Someday, many years from now, our work will be easily pinned to this decade. Today, the cloud of presentism can feel almost invisible. But in hindsight, our visualizations will have our 2020’s fingerprints all over them, just like our TV shows, just like our clothes. Inevitably, human analysis is as much an investigation into the unknown as it is an artifact of our current society. Intertwined with every new revelation resides a reflection of the moment in which the discovery was made: what we feared, what we valued, what we had most hoped to find.
Unlike science, which is so often presented as a progressive uncovering of a preexisting, observable truth, humanists know that our work is and always will be of our time.
Unlike science, which is so often presented as a progressive uncovering of a preexisting, observable truth, humanists know that our work is and always will be of our time. This acknowledgement is predicated on intellectual humility. For every blindspot we illuminate, every prejudice we strive to fix, there’s a constant recognition that we can still only see so far. And this is as true for data visualization as it is for any field in the humanities. We’ve got a lens, and we create through it. We have a responsibility to perpetually acknowledge and improve that lens. If that’s difficult to remember, add a humanist to your team. We’re good at reminding folks: there is no immunity from the present.
A humanist by training, Erin’s approach to data analytics hinges on storytelling and aesthetics. Her company, Data Dozen, is a visual analytics studio that specializes in data visualization training and Tableau consulting. Working with students and professionals alike, Erin strives to demystify data with bootcamps and consultations that make data visualization approachable, exciting, and meaningful. Erin lives with her husband Jay, their daughter Madelyn, and dog Odin in Kansas City.
