MS Data Analytics vs MS Data Science

[mcjon77]

I agree with most of this.

What I will say is that most folks I know who go for a business analytics masters are not looking to become Data Scientist. Do some achieve that? Yes. However, I think the whole purpose is to become knowledgeable enough in Data Science to have an impactful conversation with a Data Science team and then relay the information effectively to the decision makers.

I personally tell folks if they want to become a Data Scientist, pursue a MS in Statistics from NC State or Purdue since they are online programs and most folks I know are looking for an online program.

I agree with your statement regarding business analytics degrees. In fact I would go a step further and say that the vast majority of business analytics programs do not provide the adequate training for someone to become a data scientist. They don't cover statistics or the computer science portions of the field nearly enough.

However, what has to be emphasized is that business analytics degrees should be considered separate from other analytics degrees.  A program like Georgia Tech's Masters and analytics is a completely different animal. I like the program because it gives someone an very solid statistics foundation, much more so than most data science masters programs, and couples that with a very solid programming foundation along with at least two key business courses.

The reason why I like the two business courses, one being a business fundamentals course in the second being a business analytics course, is that so many data scientists just don't have a deep enough understanding of core Business principles and how their work contributes to the bottom line of the company. this kind of information is critical when deciding on what projects to take on and how to present your project to prospective stakeholders.

 While I absolutely believe that a masters in statistics is a great option for folks who want to become data scientists, the more I work in the industry and look at the curriculums of some of these applied stats programs the more inclined I am to recommend a stats Masters degree only for people who especially like statistics.  If a person likes the cs/ml aspects of data science more I would tend to recommend that they get a master's in computer science instead. UIUC has a great Masters and computer science with a concentration in data science.

Obviously, most stats masters programs get very deep into statistics and designing experiments, etc.  however, many seem to be lacking in both the cs and the business aspects of the field. A lot of stats guys dismiss those things as being something you can just learn on your own, and you can, the same way that CS folks can pick up a lot of stats on their own.

However my opinion is scarred by my experience rewriting code created by brilliant statisticians who were absolutely horrible programmers. The code was fine for one off projects, but it required a complete rewrite to use that same code for more repetitive processes, let alone to push it to production.  

This job actually has a surprisingly large amount of coding and many stats programs just don't seem to spend enough time doing that and developing clean programming and code optimization habits.  Additionally, one of the most surprising aspects of this job is that the vast majority of meetings and discussions that we have in my team of data scientists don't involve a lot of deep statistical questions. Most of them are it/cs questions.