Developing the Interdisciplinary Data Scientist of the Future

Tom Pohlmann, Head of Values and Strategy, Mu Sigma

Developing the Interdisciplinary Data Scientist of the Future

Over the years, several studies have highlighted the shortage of data science talent in the private sector. But, what is causing this shortage? Is there not enough interest in the field, and if so, is there simply not enough clarity in available job roles to spur that interest? Or, are universities losing undergraduate talent to other disciplines?

Universities and online learning platforms have responded in their own way, by offering data science curricula to generate more interest in the field among undergraduate and graduate students. In fact, according to Datascience.Community, there are more than 600 college-level programs around the world offering degrees in data science, with a vast majority of them here in the U.S. Large businesses have responded as well and are placing a premium on positions that require analytical skills and hiring data scientists in greater numbers.  A recent Datanami article described the data scientist as the number one job in the country, citing Glassdoor, with a starting salary of nearly $117,000.

What both universities and enterprises are overlooking, however, is that the real problem lies in the skill sets of the myriad business analysts and data scientists already present to perform the work at hand. This shortage is more a result of data scientists not being trained in problem solving, business acumen, or in how to translate data and analytical insights into business actions. This new, more valuable, kind of data scientist is where demand is far exceeding supply.

One of the reasons this has become a pressing issue is because many universities are still focused on building narrowly defined skills – like building the perfect multivariate regression model – instead of the interdisciplinary skills needed to solving complex business problems in areas of marketing, risk, and supply chain. In doing so, universities and businesses can develop solutions in the long term (5-10 years), medium term (3-5 years) or short term (1-2 years). Below, we’ve highlighted potential solutions for building a more robust pool of highly talented data scientists.  

Medium Term to Long Term

What’s encouraging about today’s data science field is that commercial enterprises have started working more closely with universities to develop talent and prepare them for work in the real world. However, most of these university and business collaborations still have room for improvement. When looking at most of the collaborations that currently exist, one can see that they are either limited to a particular industry – such as health care analytics – or a specific capability – like machine learning. Of course, industry relevance and in-depth exposure to emerging capabilities are essential, but moving forward, enterprises will need to continue to aid the educational system in simulating real world work environments by also including aspects of organizational behavior (behavioral/social psychology) and even design thinking (customer needs + technology + business). Universities need to make the creation of an interdisciplinary approach to developing talent at the undergraduate level a high priority to meet the demand today and tomorrow for data scientists who cannot just build the right regression model, but effectively help drive behavioral change and outcomes as a result of their work. Programs at universities like Carnegie-Mellon, NYU and University of Missouri are out front in this regard. It’s these nuanced, interdisciplinary problem solving skills which can’t be replaced by automation.

Short Term to Medium Term

In the medium and short term, organizations should focus on nurturing the skills of their existing talent in new ways, rather than constantly seeking new talent from the outside. The first tactic is to rotate hi-potential analytics professionals across functional areas within the organization on a more frequent basis. This doesn’t just mean rotating people across sales, marketing, operation and finance, but it also means rotating the actual day-to-day work and the skills that are being used. Successful organizations are also able to move people across analytical domains (e.g., reporting versus diagnostics work.), and give team members experience in being both an interpreter and producer of data analytics. Another tactics is to create interdisciplinary SWAT teams that can rapidly respond to needs. These teams are comprised of experts from various business functions and various skill sets that focus on a family of problems. Although this approach is effective in some cases, it is not a highly scalable model in the long term, which is why universities must develop interdisciplinary skills at the individual level. The third tactic is to train existing talent in interdisciplinary skills by collaborating with and learning from vendors and universities. This approach is preferred for the short term, as businesses won’t have to create talent from the bottom up, but rather focus on refining their current talent.

While this shortage in data sciences isn’t necessarily referring to the physical number of people in a given space, but rather the level of interdisciplinary skills, there are many ways for businesses and universities to make strides in building up their current talent pool. Instead of focusing purely on building data analytic skills, focusing on building interdisciplinary skills across multiple fields, whether it’s the long, medium or short term, will help improve upon our current pool of data scientists.

[Image courtesy: geralt/Pixabay]


About the Author

Tom Pohlmann is the head of values and strategy at Mu Sigma, a leading global provider of decision science and big data analytics solutions. In this role, he is responsible for brand and communications, a portfolio of new client accounts and the development of programs that continuously align the company's work with its values and vision. Tom has more than 25 years of experience in strategy and product development, messaging and branding, profits and loss accountability, mergers and acquisition, customer insights and analytics, publishing and public speaking.