Data Science: Scientific Discipline or Business Process?

Simply put, data science is an attempt to understand given data using the scientific method. That's why data science is a scientific discipline. You are free (and encouraged!) to apply data science to business use cases, just as you are encouraged to apply it to many other domains.



Woman yelling at cat

 

I'm often critical in conversation and online discussion of explanations of "data science" being intrinsically linked to business in any way.

A typical back and forth on the subject will go like this:

  • Someone will define data science in some manner which is directly tied to the benefit it can provide to business, for example "data science is what business uses to get the most out of their collected data"
  • I'll comment that data science is not about business at all, and that it is a scientific approach to understanding and gaining insight from data
  • I'm often then accused of not understanding data science, and sometimes assumed to not be very good at it; "If you were actually good at data science and were able to add value to business you would understand the connection between the two"

Like there's no way that I could possibly fathom the value that data science can add to business if I disagree that it is innately a business process.

What I can't understand is why there is such a misunderstanding of the simple idea that while data science is not inherently a business process, it is (of course) useful to a wide variety of domains, including business.

Imagine you have a degree in chemistry, making you qualified to be referred to as a chemist and to practice chemistry. Does this innately have anything to do with business? No, of course not.

But let's say you are hired by BASF. Or DuPont. Or ExxonMobil. They would then want you to use your chemistry knowledge in order to add value to their business. They would be interested in extracting the relevant know-how you posses in order to enhance business processes.

The chemist sees value in chemistry; the business sees value in the chemist using their knowledge in the scientific discipline to quantitatively increase some measurable business metrics.

Similarly, the "value" for a data scientist is whatever insight can be gleaned from some specific data. From a business perspective, the "value" which data science can provide is some quantitatively measurable business metric that is improved via this gleaned information.

Simply put, data science is an attempt to understand given data using the scientific method. That's why data science is a scientific discipline. Plain and simple. The fact that the word "science" is included in the term, while the word "business" is not, should be a dead giveaway.

You are free (and encouraged!) to apply data science to business use cases, just as you are encouraged to apply it to many other domains. But this application does not permit the wholesale co-opting of data science as an exclusively business term. And some of the biggest problems that we can find insights into using data science — weather prediction, climate change, earthquake prediction, to modestly name but a few — have nothing at all to do with business.

I understand the reaction of some to co-opt this term and claim it for the business world, given that many of data science's most prominent practical applications can be found in the business world (recommender systems, customer segmentation, decision making). But chemistry exists without the chemical processing industry, just as data science exists without business.

Data science grew out of academia and stretches back decades, pre-dating the 2012 Harvard Business Review "sexiest job of the 21st century" quote, as well as the coining of the term in the decade prior. And while this scientific discipline will undoubtedly outlive its current business buzzword-iness, it will continue to drive business decisions for decades to come.

So here's where I stand: Data science is a scientific discipline which can be applied to business processes. Feel free to disagree or express nuance below.

 
Related: