Why We Need Humanists in the Age of Big Data

Because someone has to tell us what all that data means, says former Senior Fellow James Shulman in his op-ed. To read more about the impact and outcomes of humanities education, check out the Mellon Research Forum section on our site.

TextDNA facilitates large-scale overview analysis of patterns in linguistic data sets. Humanists have distinct abilities to examine data from multiple angles while interpreting trends and outliers (Image courtesy of University of Wisconsin-Madison Graphics Group).

People who study the humanities need jobs. As the price of college to families and college debt continue to rise, subsidies from states could come with more strings of practicality attached (e.g. calls for “differentiated tuition,” with higher rates for students who study the humanities). Since the economic shocks of 2008, the humanities—under suspicion for decades if not forever as impractical—have faced evermore stern critique. Students majoring in English, history, or philosophy grow anxious when called to report at family gatherings. For thousands of students, the disdain for following humanistic paths must feel like the working world is calling them to account for their self-indulgence. They wonder where they went wrong and try to figure out how to magically summon a career translation genie. Unlike software coders or people with accounting degrees, they cannot choose where they want to live or even discern paths forward that have any degree of rhyme or reason (and these are people who know a lot about rhyme and reason).

But there are jobs out there. The talents that humanities majors have can be used for forward-looking and satisfying work. In December 2016, The McKinsey Global Institute (the research arm of the famous consulting organization) published The Age of Analytics: Competing in a Data-Driven World. The report traces how big data, data science, and analytics have changed the landscape of the workforce since McKinsey’s last report on the subject in 2011. At this point, a humanist reading this post will sigh and conclude that despite my title, this post is not for them: “I know, I know. I should be writing algorithms to find patterns in enormous pools of data; I should be coding, but I don’t know how to write software!” One prepares to hear yet another lecture about how those in the world of machine learning, artificial intelligence, and data science will be the winners in the game. 

Not so, reports McKinsey. One of the report’s key findings is that one of the biggest obstacles to making full use of the data explosion are difficulties in finding and retaining “business translators.” As it turns out, computer scientists and database administrators are not the only ones needed to unleash the potential of big data: 

Many organizations focus on the need for data scientists, assuming their presence alone will enable an analytics transformation. But another equally vital role is that of the business translator who serves as the link between analytical talent and practical applications to business questions. In some ways, this role determines where the investment ultimately pays off since it is focused on converting analytics into insights and actionable steps. In addition to being data savvy, business translators need to have deep organizational knowledge and industry or functional expertise. This enables them to ask the data science team the right questions and to derive the right insights from their findings.

This broad category of work—this much needed function in the core of the data analytics world that will re-shape every industry and every public sector activity—is crying out for the distinctive capacities of humanists. The study points out that context matters—the rich, localized, interplay between the substance of reality and the process of reducing those moments and visions of the real into data tables. Data science is far from a monolithic capture of the world; building databases and knowing what questions are sought from them requires (as the report notes) knowledge of the industry in question. There is no one data structure that defines every field of endeavor. Siloes have grown up for a reason—and while oft-times siloes within a domain can be redundant, localized data addressing particular fields have been created in response to the need for specialized structures for categorizing knowledge in countless fields of human endeavor. Everything that you use and buy and every social issue that you are concerned with will have its own data needs and its own complexities. The authority file that allows one to search for “haricot vert” on a grocery delivery site resulting in “string beans” is not the same thesaurus that allows you to navigate from “heels” to “Kitten Heels” to “2 inches” to “Prom and Homecoming” to “Taupe” on a shoe shopping site. And both are completely different again from a dataset that tracks whether there are genetic predispositions to migraines. Every industry out there, every product, every aspect of public service, every realm of human activity has its own characteristic analytic context.

What McKinsey is noting is that a few things need to come together: (1) industry knowledge; (2) data conversancy, and—while they do not say it in these words—(3) the brainpower to make an argument from evidence, i.e. the ability to see what matters in a story that is being told somewhere; the creative interpretive skills to know how the pieces of knowledge can be organized and converted into meaning. The McKinsey study is calling out the need for people who can categorize, organize, and find meaning in the colorful and flavorful subtleties of human endeavors while avoiding reductive fundamentalism—the naïve idea that every can casually be reduced to some uniform binary code. Left to their own devices, the algorithmic wizards behind the curtain do not know how to figure out what matters. Humanists do.

Sure, there is a data side of this translating work: one must be able to grasp the ways that datasets are organized and managed. Some data are best kept flat—with simple one-to-one relationships between fields (like a phone book). Other structures make more sense as complex objects, wherein various manifestations (or views) can refer back to one core record (like how a coffee table in a store might have relationships to other pieces of furniture, or other products of the same material, other inventory held in the same warehouse, or other products of the same price—each of those relationships would be captured in tables in other databases.) Resource Description Frameworks (or RDF) may or may not make sense when there are multiple relationships between data fields. In assembling data, the data translator must be able to understand the work (and the intentions) of the existing data—why was it gathered, how was it structured, whether the fields were controlled or people could enter whatever they wanted, how far back the data were gathered in the same way. Often one huge barrier when working with legacy datasets concerns decisions about which fields “map” to others, which details can be lost in this process of extracting the data from one database, transforming them so that they can be loaded into another, and then loaded in a way that does not break the database design (or end up not making sense). This happens daily when companies are merged or acquired and each company’s data needs to be combined into the main stream. But this all can be learned without going back to school.

What the McKinsey report shows is that the work of making the data meaningful cannot happen without the sensitive eye of the business translator. These translators need to be like people who live on the borders of two countries and, by necessity, speak the language of both in order to go about their daily business. Harvard Economist David Deming has also published research that shows that the labor market is increasingly rewarding social skills. And beyond knowing how to have these conversations, people on the borders are sensitive to the ethics—the norms, mores, and values of the cultures involved. Humanists can do this, and it matters.

But how does someone go from Hegel to shoes, string beans, or jet fuel? You go there—as Jon Stewart suggested in a graduation speech—by finding something you love and then “getting good at it.” If it is absolutely and incontrovertibly your personal truth that all you can care about is Nominalism or Ronsard or the Brazilian Constitution of 1824, then maybe a PhD program and academia calls you. But if you care about something—about housing policy, health care, car engines, iron smelting, white label adhesives, immunizations, French fries, or support services for farm laborers—you are needed, now!—for your ability to understand and communicate about the data that field is now and will continue to both generate and need in order to thrive. 

The friendly critic of academia, Louis Menand, has argued that “the divorce between liberalism and professionalism as educational missions rest on a superstition: that the practical is the enemy of the true. This is nonsense.” It is nonsense to believe that people who get a degree do not want or need meaningful work in the real world. An academic major gives a person a set of lenses and tools for looking at the world. As the ongoing digital revolution transforms work in that world, the flexibility of mind and the ability to synthesize and to derive meaning from evidence—often praised as the aims of any liberal arts program—are (according to the world’s most famous consulting group) needed in every field. In Pragmatism, William James argued that words are not magical—they are useful only as they are applied and worked through: “But if you follow the pragmatic method, you cannot look on any such word as closing your quest. You must bring out of each word its practical cash-value, see it at work within the stream of your experience.” Arch-purists in the humanities may shudder. But the good news is that humanistic skills—when worked into the “stream of …experience” in the real world—are needed. One should neither throw away what one has learned in the humanities, nor should you berate yourself for not writing Python, constructing diagrams of neural networks, or majoring in plastics. There is plenty of hope. Having a pathway to cash-value is a pretty good thing.