If You Want to Become a Data Scientist, Skip the Computer Science Degree

Save yourself the time and money, and do this instead.

Photo by ready made from Pexels

Whether you want to be a data scientist, a software engineer, or even a web developer for that matter, you should break with tradition and skip the traditional 4-year computer science degree.

With data science being one of the sexiest fields to get into (thanks to nearly guaranteed job security, the ability to work from anywhere in the world, and the notable salaries), it may seem counterintuitive to take your journey in a different direction. But trust me, thinking outside of the box might just make you an even more desirable candidate. Don’t just take it from me though, let’s hear what the data has to say about it.

Why skip the computer science degree.

• Traditional computer science degrees can be out of date and can be at least 10 years behind in a field that changes every 10 minutes (programming languages such as Rust, Dart, Kotlin, Julia, TypeScript, and Swift, and data science necessities, such as Jupyter Notebooks, TensorFlow, and many others, were all created in the last 10 years, whereas the average computer science degree is focused on programming in C++, a language from the early ‘80s).
• Graduates from a computer science program have mostly theoretical knowledge but not as many practical skills that employers need when they graduate (ex. working with a development team, or applying programming principles to real-world problems).
• Computer science degrees generally don’t have courses or technologies that employers are looking for (ex. iPhone and Android development).
• Little time is actually spent learning how to code in computer science degrees, with a majority of the courses focused on mathematics, systems, and computer science theory.

The data says…

*A quick disclaimer: a portion of the data discussed in this article was initially collected by Indeed who analyzed tens of thousands of resumes submitted by current and former data scientists to their platform. Because the subjects are voluntarily submitting their resumes, the study completed by Indeed is potentially subject to self-selection bias, voluntary response bias, and response bias. However, because tens of thousands of resumes were used for their study, a rough generalization of the backgrounds of data scientists can be understood to be relatively (though by no means perfectly) representative of the population.

A study conducted by Indeed in 2018 offers some insight into the educational and occupational backgrounds of data scientists. While the study separately analyzes the individual roles of data scientists, data analysts, and data engineers, I find that in the general data science community there is a great degree of overlap in the definitions of the roles. Therefore I will seek to offer a “sub-analysis” of the Indeed study that represents the results of the three professions combined in an attempt to draw new conclusions that may be of use or interest.

Highest level of education achieved:

Of the five different professions Indeed analyzed (data scientist, machine learning engineer, software engineer, data analyst, and data engineer), data scientists hold the highest average level of education, with over 75% of data scientists having advanced degrees (Masters or PhDs).

It’s important to note though, that a Ph.D. is not a requirement to become a data scientist, despite 20% of data scientists in the study holding that accreditation.

Another analysis (not conducted by Indeed) of the education levels of data scientists reveals similar results, showing that roughly 44% of data scientists hold Master’s degrees, while approximately 21% hold Bachelor’s degrees.

However, if you look at the more general overlap of the data scientist-data analyst-data engineer professional area as a whole, the distribution of credentials is much more balanced. Over half of data analysts have a Bachelor’s degree, and a surprising percentage of data engineers (just slightly less than 20%) only have a high school diploma. Both the data analyst and data engineer professions also have the highest representative percentages of individuals who hold associate’s degrees as their highest level of education achieved.

Field of study:

This is where things get really interesting.

According to the Indeed study, of the five professions studied, data scientists exhibited by far the greatest diversity in terms of post-secondary fields of study.

The fields of study completed by data scientists included computer science, engineering, business/economics, math/statistics, natural sciences, data sciences, social sciences/liberal arts, and many others.

Most surprisingly, is that compared to the four other professions, the range of disciplines studied by data scientists was much more diverse. Compared to the software engineer role which was dominated by a >50% study focus on computer science, it appears that data scientists come from all different disciplines of post-secondary education.

When looking at the general overlap of the data scientist-data analyst-data engineer profession as a whole, there is a wide distribution of fields of academic study, with the three data-related professions carrying the largest Gini Impurities of the five professions studied (a large Gini Impurity indicates greater diversity in the fields of study). Data scientists came out on top with 85%, followed by data engineers with 79%, and data analysts with 78%. In comparison, software engineers only had a Gini Impurity of 53%.

If not a computer science degree, then what?

Based on the data discussed above, some general conclusions can be made about which education path will be most beneficial if you want to become a data scientist.

• Level of education: While it’s possible to become a data scientist without any advanced degrees, it’s relatively safe to say that at least a Bachelor’s or a Master's degree would give you an obvious leg up against others in the field. The data science field is in many ways ahead of its time, but I also believe that the field may still believe in the importance of higher education. Therefore, while I’m suggesting that you shouldn’t spend your time on a 4-year computer science degree, I’m not suggesting that higher education isn’t necessary for the field — if anything, it’s a great indicator of the likelihood you’ll get hired. This leads me to my next point…
• Field of study: Now, to answer the question you’ve all been waiting for: if not computer science, what should I major in? The answer: anything. If the data discussed above has taught us anything, it’s that data scientists come from one of the broadest ranges of academic disciplines seen by any industry. Furthermore, with 50% of the data science jobs coming from fields that aren’t tech-related, there’s a good chance that having some industry-specific knowledge can give you an advantage against other applicants. Therefore, study whatever interests you. Major in engineering, the natural sciences, mathematics, liberal arts, or business — anything you like really!

How to get the requisite data science and programming skills:

The beauty of becoming a data scientist is that the skills required for the profession are obtainable without spending all your hard-earned money on a computer science degree. Therefore, while you are spending money on a degree in a different academic discipline, you have some options for getting the necessary data science skills:

• Complete an associate’s degree or university certificate in software development, software engineering, or computer science. If you’re looking to learn programming and data science-related skills in a formal academic environment, an associate’s degree or university certificate is the perfect option. Associate’s degrees and university certificates in tech-related disciplines are ideal because they can be completed in as little as one to two years (one year for certificates, two years for associate’s degrees), and the curriculums often contain courses reflecting what the industry is looking for in graduates — making you an obvious hire. Furthermore, associate’s degrees and university certificates are more able to quickly change in response to changing trends in tech, so you’re not wasting any time learning out-of-date technologies. When I was first at university, I completed an associate’s degree in software engineering. The program was perfect because it offered the perfect mix of hard and soft skills that were sought after by employers in the area. The focused degree gave me all of the practical skills needed to join a company and hit the ground running, writing code, and participating in software development immediately.
• Build your own learning curriculum. This is by far the cheapest option out there, although arguably the toughest. Thanks to the wealth of information available on the internet, it’s possible to teach yourself everything you need to know to become a data scientist. From programming to mathematics and statistics, to data analysis and machine learning, there is a resource out there for it all. However, be warned: MOOCs and online courses alone are not guarantees of a job in data science. While other tech professions (such as software engineering or web development) are home to many individuals with no formal academic qualifications, data science is slightly different in that it still appears to firmly prefers candidates with some form of post-secondary education. Those that can enter data science with just an education from MOOCs are the exception, not the rule. Therefore, it’s important to think of your data science learning curriculum as a complement to some form of formal post-secondary education.

What about an advanced degree?

A search on Indeed for data scientist jobs in Canada yielded hundreds of ads, each with different qualification requirements. Overwhelmingly, the ads generally leaned towards requiring at least a Bachelor’s or Masters, with few ads requesting PhDs.

From this quick job search alone, it’s easy to see that a Ph.D. is not a requirement for a job in data science. This is further supported by this article which perfectly describes how a Ph.D. is only necessary if you intend on working for the likes of Google or Airbnb. The article goes on to describe how a Ph.D. takes an incredibly long amount of time such that those years would be better spent actually getting experience and working in the data science field.

When it comes to a Master’s degree, this article further explains that it may or may not be a good idea for you to get one under your belt depending on your unique situation (I specifically suggest you go and read the article because it offers a lot of great insight). However, if you’re dead set on getting a Master’s to kickstart or further your career in data science, you’re in luck. Most top-flight universities (including MIT, Stanford, Carnegie Mellon, University of Toronto, University of Helsinki, University of Calgary, and many others) have Master’s degrees specifically tailored around data science.

Final thoughts.

As unconventional as this advice may seem, in today’s day and age, it often pays to think outside of the box. Fortune tends to favor the brave, so breaking with tradition isn’t to be frowned upon — especially when you’re looking to get into a field as dynamic as data science. The beauty of foregoing the traditional 4-year computer science degree in favor of something a little different? You’ll be in good company.

One of the best things about data science is that you’ll be working with people from all sorts of different backgrounds. It would be boring to be working with people who all hold the exact same educational background. Instead, the beauty of data science comes from the colorful patchwork quilt of identities who all bring unique perspectives and ways of thinking to the table. In other words, be unique, play to your strengths, and you’ll be guaranteed a job in the current data goldrush.