Choosing the Right Graduate Degree for Data Science

Author: Jonathan Nolis
6 / 23 / 2017

I’m often asked the same question from people who are interested in analytics and data science: “what graduate degree would be the most useful to me?” There are lots of choices of degrees, many that claim to prepare you for the field, but different programs teach you drastically different skills. These differences can be meaningful when applying to analytics jobs and entering the industry. Here are my opinions on the different options, having worked with and interviewed many data scientists over the years:

An MS in applied mathematics or statistics degree

These degrees are good for critical thinking, but you’re unlikely to learn that many practical skills. You’ll take a lot of courses like real analysis that train you to think hard, but don’t teach you many practicalities. At the end of the program you’ll be an expert at looking at numbers and trying to understand them, but you are probably not going to know much about how to write robust code, or how to deal with messy data. Fortunately understanding the theoretical background should make learning the practicalities of data science easy, but you’ll have to do it on your own time. Since most of data science is writing code and dealing with messy data, employers will be skeptical that you can transition from academia but will trust that you are smart enough to do it once you arrive.

Recommendation: you can pursue one of these degrees (I did), but make sure you supplement it with practical, project-based work like an internship or a personal data science project on the side.

An MBA with an analytics focus

This will teach you a lot about how businesses work, but very little about how to do analytics. These programs tend to be very heavy in Excel and expect that people are coming in with little technical background. You don’t get deep experience with complex data or programming, and unlike a math/stats degree you won’t even have the theoretical background to understand the techniques. If a program can claim to teach you analytics over two years while barely moving out outside of Excel, they are doing you a disservice.

Recommendation: avoid these programs like the plague.

An MS degree in computer science with a machine learning or AI focus

These are great, as they will teach you a lot of the technical skills and you will learn how to implement them. Depending on your job you may never need to write your own algorithms (and instead use other people’s packages) but learning their inner workings makes it easy to understand how to apply them. The risk of these programs is you won’t get enough business expertise, which makes some analytics jobs harder. For example, the kind of work we do at Lenati requires applying data science and using it to make a business decisions. These degrees won’t prepare you at all for the business decision component.

Recommendation: these are good degrees, but best for people who want to be writing machine learning models more than using data science to make high-level business decisions.

An MS in Data Science

Many schools are now offering Masters degrees in data science. These degrees combine computer science, math and statistics, and business courses. The verdict is still out on how useful they are because the programs are so new, but they seem potentially promising. If you are looking at a program like this make sure multiple departments run it together. If the degree falls solely in the math, CS, or business school it probably doesn’t have a good combination of material.

Recommendation: consider them, but make sure you do extra diligence on the program material.

But regardless of what program you choose make sure along the way you get lots of project work and as many internships as possible. More than the degree itself, what matters to employers is that you have real experience working with messy data, applying models, and understanding the business context. No set of coursework can compare to an internship or even a side Kaggle competition to help employers see evidence of your understanding of real data science.

Share on FacebookTweet about this on TwitterShare on Google+Share on StumbleUponShare on LinkedInEmail this to someone

About the Author

Jonathan Nolis is the Director of Insights & Analytics at Lenati, and is the lead of the Customer Insights & Analytics team. He has over a decade of experience in solving business problems using data science. Jonathan has provided insights and strategic advice in industries such as retail, manufacturing, aerospace, health care, and e-commerce. Jonathan helps create proprietary technology for Lenati including the Loyalty Program ROI Simulator – a tool that uses big data to predict the value of a loyalty program. He has a PhD in industrial engineering, and has several academic publications in the field of applied optimization. Prior to joining Lenati, Jonathan was a Lead of Advanced Analytics at Promontory Financial Group, a regulatory compliance consulting firm.

Leave a Reply

Your email address will not be published. Required fields are marked *