Daniel Gutierrez
In this special feature, Daniel D. Gutierrez, Managing Editor of insideBIGDATA discusses a question on many people’s minds these days – is an advanced degree required to become a data scientist and if so, what programs are available? In addition to being a big data journalist, Daniel is also a practicing data scientist, educator and sits on a number of advisory boards for various start-up companies.
As the big data industry continues its incredible upward trajectory, a veritable flood of enlistees are wondering how they can get the training and education necessary to take a foothold in this exciting field. Many are coming in laterally from loosely affiliated areas such as economics and finance, social sciences, as well as physical sciences such as physics, astronomy, even oceanography. The common thread with all these future data scientists is that they need to quickly get up to speed with technical disciplines heavy in prerequisites.
The MOOC or Formal Program?
I’m fielding more and more questions over on Quora about this topic. Just the other day someone posed the question about data science educational resources. I suggested the Johns Hopkins data science certificate program offered on the Coursera MOOC platform since I had beta tested much of the course content before it went live and I served many times as a Community TA for the Practical Data Science course in the series of 9 classes. A reply to my comment came in saying that this series was light on important material, just mentioning important concepts but not providing enough detail.
I could see the gentleman’s point and used the machine learning method known as “boosting” as an example. There are books and research papers written about this topic so it is unrealistic to expect a deep level of understanding after taking a 4 week online course. MOOC offerings should be viewed as more a template for learning so if you’re a self-starter and motivated to learn, you would need to take this template and drill down into specific areas for deeper understanding.
Often however, a prospective data science professional requires a more formalized educational experience. There is huge reward awaiting the candidate who invests time and money in a traditional masters or Ph.D. degree. By engaging in such a program, you’d get the depth in knowledge the MOOC critic was seeking.
But there’s an obvious trade-off. You can get up to speed quickly with one or more MOOC resource, but much more knowledge and experience will be required to take advantage of the many employment opportunities out there. On the flip side, treading through a 2 year master’s program, and much longer for a Ph.D. means you’re out of the action in a very hot industry. Who knows what your opportunities will be and in what shape the industry will be after you get your diploma.
The Myth of the Unicorn
Another important issue with people wanting to take part in the big data maelstrom as a profession is what type of role you desire. Many job ads are for “data scientist” where the list of required skill-sets is a mile long. In reality, you need to make an early decision whether to become a true data scientist, or a big data engineer. The talent and experience areas as well as educational background are quite different for each.
In the case of a true data scientist, a background in computer science, mathematical statistics, probability theory and machine learning is the norm. You should have a talent for data modeling, algorithm selection and validation, as well as data visualization with a proclivity for data story telling.
For a big data engineer, the skill-sets are more toward the engineering side of the equation with an emphasis on software engineering, hardware architectures, parallel processing systems like Hadoop and Spark, along with system administration, scalability, data governance, security, etc.
Typically, these two classes of professionals work closely together to create production systems. Of course, some companies believe that a single person, known as a unicorn, can handle all of the above but this is not very likely. In reality, a data science/big data “team” is what’s needed although many companies resist this notion before a team requires a bigger financial investment.
Let’s turn our attention to outline some advanced degree programs designed to kick start your career in data-centric fields. These apply mainly to the true data science role.
Masters Programs
There are a variety of quality Masters-level programs from leading universities that will give you a definite leg up on other candidates applying for jobs in the field. Here’s a short list of schools offering data science educational programs. Other schools are launching new masters programs all the time:
Northwestern University – Master of Science in Predictive Analytics
UC Berkeley – Master of Information and Data Science (MIDS)
Stanford University – Master of Science in Statistics: Data Science
Columbia University – Master of Science in Data Science
NYU – Master of Science in Data Science
University of Wisconsin Data Science Master’s Degree
Ph.D. Programs
Doctoral programs specifically based on data science are a new phenomenon, although Worcester Polytechnic Institute (WPI) recently announced what they deem the nation’s first interdisciplinary PhD program in the field beginning this fall.
In order to get a doctorate degree pertinent for the field of data science, one path to success is to find a top school with highly-rated departments in computer science, statistics, or mathematics. An advanced degree in any of the above disciplines will go a long way for data science employment opportunities. To many companies, especially well-funded tech start-ups, hiring a Ph.D. data scientist is a rite of passage for technological superiority. You’ll see many Chief Data Scientists, Chief Data Officers, and Chief Technology Officiers with this level of education.
Here is a short list of the top Ph.D. programs in data science related fields as reported by the latest U.S. News & World Report academic rankings. The order of the list is based both on the rankings any my own opinion of the schools. Note some U.S. News rankings were reported as ties.
Stanford University (ranked #1 in CS, #5 in Math, #1 in Statistics)
UC Berkeley (ranked #1 in CS, #3 in Math, #2 in Statistics)
MIT (ranked #1 in CS, #1 in Math)
UCLA (ranked #13 in CS, #7 in Math, #30 in Statistics)
Harvard University (ranked #18 in CS, #3 in Math, #3 in Statistics)
Caltech (ranked #11 in CS, #7 in Math)
Princeton (ranked #8 in CS, #2 in Math)
Columbia (ranked #15 in CS, #9 in Math, #20 in Statistics)
Yale (ranked #20 in CS, #9 in Math, #34 in Statistics)
Carnegie Mellon (ranked #1 in CS, #34 in Math, #9 in Statistics),
Cornell (ranked #6 in CS, #13 in Math)
Although getting a coveted Ph.D. in any of these fields certainly will seal your lucrative career path long term, it won’t be without significant sacrifice – these programs will be very expensive and require several years of your life at minimum with full-time effort.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind