Sign in
Google (or Bing if you please) “Master’s in Data Science” these days and you will see at least four ads for Data Science degree programs.
You’ll also probably see several links to https://www.mastersindatascience.org including one sporting their ad:
“How to Become a Data Scientist: Master’s Degree in Data Science”
You’ll see articles ranking and discussing the merits of these diploma programs, as well as blogs discouraging analytically savvy folks from bothering with such a degree.
So, what’s this all about? A recent Burning Glass report, “The Quant Crunch”, found that 42% of Data Scientist positions require a graduate degree.
Analytics executive recruiters Burtch Works suggest that 88% of Data Scientists have at least a Master’s Degree.
The Quant Crunch report explains,
“Demand for a new breed of professionals skilled in data, analytics, machine learning, and artificial intelligence requires a requisite response from both higher education and workforce development.”
The Master’s in Data Science trend is just that: higher education’s response to demand – and the classic Data Scientist shortage we’ve all heard about.
Suddenly, Data Scientists and Analysts who have been working several years in the field with “just” a bachelor’s degree in Computer Science or Statistics are competing with people who have risen to the top of your candidate list with a Master’s in Data Science and higher data literacy skills.
If you are a recruiter for Data Science and Analytics roles, it can be confusing to determine whether a candidate with a Master’s in Data Science merits more, less, or the same attention as one without.
Is there substance, or even a need, behind graduates with Data Science degrees?
Or is this just a result of more hype around the Data Scientist unicorn?
A Long Long Time Ago in a Galaxy Called 2014…
Back in 2014, when only a handful of universities offered a degree named something along the lines of a “Master’s of Science in Business Analytics”, KDNuggets started a thread called “Do you need a Master’s Degree to become a Data Scientist?”
The general consensus was that good practical skills could take the place of a Master’s Degree.
Respondents to the thread commented,
“You can learn data science anywhere. No single Master’s Program{sic} could cover all the disciplines needed in significant depth.”
On the other hand, one IBM Data Maestro opined,
“To learn data science — absolutely.”
In 2014 most people with Data Science jobs did not have a Master’s in Data Science.
Lots of Data Scientists had degrees in Math and Statistics, but others with a good grasp of problem-solving had moved into the space from fields such as Economics and even liberal arts.
So, four years and dozens of new Master’s in Data Science degree programs later, is the industry opinion any different?
Let’s look at the pros and cons of a Master’s in Data Science in light of today’s Data Science needs.
Why Look for a Master’s in Data Science?
Got skills?
You do if you have a Master’s in Data Science diploma.
Dr. Daniel Wu, Coordinator for the Master’s of Data Science program at Cabrini University, attributes the need for a Master’s in Data Science to the fact that “Hiring managers are clearly struggling to find applicants who check all the boxes.”
He and other experts argue that a Master’s Degree can help students develop skills in a large number of requisite areas that some self-taught candidates may miss out on. We concur.
The demand for Master’s in Data Science programs is a direct result of the continued shortage of widespread, consistent, and reliable Data Science skills among candidates.
A diploma in Data Science indicates that a candidate has had systematic and well-defined training on both the theory and application of analytics.
A Data Science degree program builds upon specific skills acquired over time, and in the end, should impart a holistic set of skills.
This is probably the single most useful feature of a Master’s Degree. A degree provides a demonstrated commitment to and methodical learning in the field.
Practical Experience
Master’s candidates often graduate with relevant work experience.
This might be in the form of a prerequisite to the Master’s program, internships, hackathons and competitions, or reality-based projects and coursework.
Top programs, such as Northwestern and Carnegie Mellon, provide internships.
Others such as Columbia University, University of San Francisco, and even the University of Wisconsin’s online degree, provide capstone projects involving research and opportunities to interact with industry players through faculty and coursework.
Graduates with work experience are more likely to come from a full-time or two-year degree program in urban areas.
However, a Master’s in Data Science Doesn’t Guarantee
Consistent, Up to Date Learning
As with any new product, quality is not guaranteed.
Apart from the obvious winners like Stanford and Harvard, be wary of programs that are only repackaging material from other courses, that treat Data Science as the same old data mining stuff, or which mostly teach theory.
In the early days of a Data Science degree program, the rush to be relevant could mean that some academic departments have had little time to put together their Data Science curriculum.
Budgets, best practices research, and specialized faculty searches may be quickly cobbled together.
In some cases, the degree coursework might be taken from existing courses on campus. As such, the resulting coursework may or may not be up to date and likely won’t be a high standard.
Master’s in Data Science programs are also housed in different ways across the university spectrum.
Some programs have a more vertical focus, such as Healthcare Analytics. Others, such as “Information Systems”, are broader in scope.
Below is a fascinating “constellation” of Data Science programs across the USA taken from www.mastersindatascience.org.
What is interesting to note is that the blue nodes are the colleges within which the Data Science Master’s program is housed. Most are housed in the business college, but not all.
The college in which the degree program lives affects the prioritization and interpretation of what will be taught in a Master’s in Data Science program.
Source: https://www.mastersindatascience.org/blog/picking-a-data-science-program/
Critical Skills
A Master’s in Data Science certainly indicates a significant level of technical skills, but other areas of competency are just as important to being a top-notch Data Scientist.
Most business problems that need solving are bespoke by nature. They require skills such as intuition about data, critical thinking, problem-solving, adaptability, self-discipline, and communication.
These are just a few skills not specifically taught in a Data Science Master’s program.
At QuantHub, we find that what really separates people who are successful from those who aren’t, are these soft skills.
Specialized Skills
Burtch Works says there is “a distinct trend towards specialists instead of generalists” in Data Science.
They have identified four key areas of growing specialization:
– natural language processing
– computer vision and image processing
– ad tech
– the internet of things
Some Master’s programs offer specializations, but not necessarily in these areas.
A Killer Portfolio
It’s essential if you are recruiting Data Scientists at a higher level that they demonstrate what they are capable of.
Attending classes for 20 months does not guarantee that someone has put themselves out there, sharpening their skills with other Data Scientists.
A portfolio of projects on a personal website, participation and success in Kaggle, CrowdAnalyti x and DrivenData competitions, and Github open source contributions are all highly valuable experiences outside of the classroom.
These demonstrate aptitude and a commitment to learning about Data Science.
The Bottom Line
We live an age where technology makes all matter of education and work experience paths possible. We also live in the age of talent. Anyone with it can learn and achieve.
Data Science is still a rapidly evolving field. Until norms are more established, it’s unlikely that every stellar Data Scientist will be following the same path.
Data Science skills are a spectrum. At one end there is a Ph.D. or a Master’s in Data Science and at the other, the DIY Data Scientist who took a couple of statistics courses online.
There are people with a B.S. in Computer Science that have been working in Analytics for over ten years. They can probably do just as much, if not more, than many higher education graduates.
There is plenty of space in the middle of this spectrum to find the right candidate.
So, do your recruits need to have a Master’s in Data Science to be good at it? Although it won’t hurt, we think not.
At QuantHub, we do believe one thing to be true.
Master’s Degree or not, employers should proactively evaluate and measure a Data Scientist’s skills – these skills aren’t a given.
Candidates should be able to demonstrate their aptitude outside of a classroom. So, our recommendation would be to focus on aptitude and demonstrated ability to do the job vs. academic knowledge.
We’d love to hear your thoughts, so hit us up with comments below. Happy hunting!
About the Author
Matt Cowell is the CEO at QuantHub, an AI-driven platform for attracting, vetting, and developing data scientists. QuantHub helps recruiters and corporations vet Data Scientists and related Analytics professionals to truly gauge their level of expertise. QuantHub’s comprehensive evaluation platform covers skill tests (Python, R, Statistics, etc) as well as real-world data challenges to verify that candidates have both the skills to do the job and also have the ability to apply those skills to real-world responsibilities. Visit QuantHub.com today to find out how you can begin to identify and vet the best Data Science candidates!