Your analytics team needs a liberal arts major

Here’s a modern cliché: Data science and analysis is an engineer’s discipline. But it’s wrong. LinkedIn tells us that liberal arts graduates are joining the tech workforce at a faster rate than technical graduates. Scott Hartley’s book, The Fuzzie and the Techy (2017), outlines how liberal arts skills are vital in the modern tech sector. Whether it’s anthropologists working at Nissan or sociologists at Tinder, successful technology requires skills from the arts as well as from engineering.

At the same time, governments are cutting funding for the arts to focus on STEM (science, technology, engineering, and math). This is a problem and we need to bring the arts back into the mix to build a strong pipeline of multiskilled critical thinkers for the jobs of the future.

When I was 16, I wanted to specialize in English, math, and art, but I was told it was a terrible choice because it was too diverse. It took me 30 years to realize the error of that advice and I now realize my success in business intelligence and data analytics has been enabled by diverse skills.

Liberal arts skills are needed at every level of data science, from storytelling and communication of insights through to the training and interpreting of the most advanced machine learning applications in your business. Here’s why.

Michael Corell, research scientist at Tableau, says, “People are involved in every step of data analysis: from collection to storage, analysis, and decisions. Traditional BI often makes these people invisible, which can have dire consequences for the responsible use of data. The liberal arts are a way of making these people visible again, and promoting better, more human decisions.”

One example is how charts eliminate any human emotion and meaning from the information, discussed in detail by Sam Dragga and Dan Voss in “The Inhumanity of Technical Illustrations.” They considered how charts reduce human tragedy (for example, data about deaths in industry) to unemotional data points, and question whether that is appropriate.

Ben Jones, author of the DataRemixed blog, agrees. “Communicating data requires people to think of themselves as more than just an engineer,” he says. “Insights need to be communicated. Successful communication requires design, journalism, and artistic skills.”

It is not just in the presentation of data end that we need skills from the liberal arts. As machine learning impacts more of our lives, how do we ensure that that data being used to train the algorithms is good enough to guarantee fair outputs? Caroline Sinders argues that the most important job in the future will be in data ethnography. “This is the study of the data that feeds technology,” she says. “It will consider data from cultural perspectives as well as data science perspectives.”

We’re encountering many situations where poorly implemented machine learning algorithms create problems. For example, algorithms that discriminate against poor people, teachers being graded incorrectly, and search algorithms identifying black people as gorillas. Ensuring culturally diverse assessments of these algorithms could well avoid the problems.

As you build an analytics group in your organization, you need to compliment engineering skills with liberal arts skills. A focus on specialization over many decades has led to people getting siloed within disciplines. CP Snow, a leading English chemist, lamented this schism way back in the 1959 in “The Two Cultures.” He described how these siloes harm policy, education and culture itself.

We can address this by training people differently. An example is the Data School in the UK, run by Andy Kriebel. “I’m of the opinion that people with liberal arts background are at an advantage because they’re more likely to have done some creative studies,” he says. “People with engineering background tend to be much more prescriptive and go into way too much depth without focusing on the consumer.”

Whether you are a company director, an educator, or a parent, you need to think about how to create a diverse workforce in the short and long term, equipped to manage the data that has an ever-increasing impact on our society.

This article is published as part of the IDG Contributor Network. Want to Join?