In Defense of PhDs for Careers in Data Science

“Skip the PhD and Learn Spark, Data Science Salary Survey Says,” claimed a recent report.

On the surface, the data appears to make a case for this approach. Spark, or indeed any technology making waves right now, probably represents a very good way to get in-demand (and therefore well paid) roles.

But as someone for whom a PhD has worked very well, I want to put the counter argument for their value in the world of data analytics. One that looks to the longer game.

Firstly we should mention that PhDs are not purely about financial value. It’s certainly true that the majority are done based on a genuine desire for advanced work in that subject – PhDs are hard to complete without that personal drive. The impact on earning power is often a secondary consideration, if taken into account at all. Money is not the primary driver.

But let’s assume you are looking at a data science or related PhD as a career move, and considering whether learning specific skills might be a better option.

The survey highlighted Spark as having a particularly strong salary boosting effect, which looked at from one analysis appears to be higher than that of a PhD. This, alongside other data in the report, paints a picture of the ‘in vogue’ technologies with the highest current market value.

But be careful how you read this. Spark may be the hot ticket in 2015, but how long will it last? Maybe there’s a demand bubble now. It might be gone by 2017, to be replaced by a bubble for Frizzle (a technology I made up). Then what? Has your market value peaked, to be quickly eroded by cheap whizz kids fast arriving on your heels, full of new Frizzle skills gained whilst you were still Sparking away?

It’s important that you create a clear picture of the kind of “data science” career you want. Data science makes the most a difference when the analytics team sits real close to the business units that need the derived insights. Those data science teams need a rich blend of raw technical, interpersonal, problem solving and business domain skills.

Turning up with a skill in a technology makes it more likely that you’ll be solving a problem that’s already been defined for you. And you may find it difficult to make the move that makes you the problem definer and solution designer. In other words, the fun, high value bit. You might be a bit tired of Spark by the time that happens. Data science can offer you a lot more, if you offer it a broader set of skills.

That’s not to say Spark isn’t a valuable thing to learn, and there is evidence it will boost your earnings in the short term. Equally PhDs are expensive and are not right for everyone.  But if you are looking for a career, rather than a short term dash for cash – you need to look at specific technologies as being just part of a broader skill set.

My point is, your future is important, so think hard about it. Try and look at a career profile over a longer term, where the skills you bring actually grow in value rather than diminish at the speed of technology refresh.

About the author: Dr Nick Clarke is head of analytics at Tessella, an international analytics, software and consulting services company.  His PhD is in Quantum Chemistry.

Related Items:

Skip the Ph.D and Learn Spark, Data Science Salary Survey Says

10 Tips for Beginning Your Big Data Journey

Data Science Education Gets Stronger, But It’s Not There Yet