Should you take more online courses or master's to boost your resume for data science jobs? 3 scenarios and recommendations

While speaking to aspiring data scientists, I often get asked: “Should I take more online courses to boost my resume?” This article summarizes my personal recommendations, and how to identify when one is “ready” after self-learning or pursuing additional education.

This article may help you if…

Throughout mentoring aspiring data scientists, as well as speaking to current data science practitioners (in industry and academia), a common question is how one can best self-learn machine learning from online platforms such as Coursera, Udacity, or edX, to name a few.

As an added layer, those that do use these platforms to self-learn, run into a decision point after they have completed an online course or two: “Should I take more courses, or is this enough?” A similar and common question is “Should I get a master’s degree in data science / CS / statistics?”

One last common thread, is that even after taking online courses and self-learning, a lot of it seems to be forgotten just as quickly as a midterm exam in university. I’ve encountered this too: “Didn’t I just take that Coursera course on CNNs? This [concept] rings a bell, but how come I forgot the details already?” This is a horrible realization to have in general, but imagine if you were in a job interview!

If any of the above situations resonate with you, I hope this article can help. From years of personal experience, I share what can help you identify when you can stop with online courses and know that you are “ready”.

Rule of thumb: Diminishing marginal returns

This rule holds true for self-education with online courses, or additional formal education such as getting a master’s degree.

Diminishing marginal returns is a concept from economics: Imagine a brand new restaurant getting ready to open for business. They just spent money on a brand new kitchen. When they have no cooks, they have no output, and thus, no customers or returns, no matter how amazing a kitchen they have. When they hire the first cook, the output and returns increases a lot, since they go from serving zero customers, to being able to serve their customers.

Then, the restaurant gets more popular, and the one cook isn’t able to handle more orders. A second cook is hired. This increases the output and the returns. However, once past a certain point, maybe when they have 3 cooks, hiring more cooks isn’t going to further increase returns. The kitchen could get too crowded, making each cook’s output decrease! At this point, the restaurant shouldn’t hire more cooks, but rather expand in other ways, such as making the kitchen larger.

This analogy applies to online courses and additional education. If you have zero of it, doing an online course or two will have incredibly high marginal benefit. Past a certain point, you need to look at expanding your skills in other ways, such as doing a side project - akin to expanding the kitchen in the restaurant example, instead of hiring more cooks.

Why are you taking online courses, anyway?

Now that I’ve gone through the general rule of thumb, I want to help you apply this logic to your own unique situation. Since there are many types of people that are looking to improve their skills and credibility in the data science field, I want to give you the tools to make the decision based on your own educational background and work experience. Below, I list several broad demographics for you to start with.

The most common reason and demographic that look to online courses or additional education I’ve seen are aspiring data scientists. Taking online courses and getting certificates of completion is something to put on a resume in order to get an interview, which if the candidate can pass, lands the candidate a job in the data science or machine learning industry.

I’ll break this demographic down into two types: Type 1 are those that don’t already have any job experience, and are fresh out of school or from a bootcamp. Type 2 are professionals who do have job experience in other fields (academia included). They want to transition or pivot into the data science field.

The next demographic, Type 3, are existing data scientists looking to expand their knowledge or enter a new specialization (for example, from NLP to reinforcement learning).

One last type are those that just like to learn for fun, without any defined reason. They don’t see online courses as a way to get somewhere else, but simply entertainment or mental fulfillment. If this is you, then you should stay true to yourself and keep taking online courses, as long as you’re having fun!

Types 1, 2, 3 all have something in common: They have a place they are in the present, and a goalpost (data science/ML job or promotion) they want to move to in the future. They hope that online courses can fill in the gaps to take them from point A to point B.

The 3 types of people that could benefit from online courses or additional education

We’ve established that no matter which type you are, you aim to use online courses or additional education to reach point B, which is your goalpost.

I will walk through the following 3 examples, but my intention is that you can then apply the logic to your personal situation:

Step by step analysis of the 3 types

[Type 1] “I just graduated from university, but I hear that I need a master’s to get into data science. Should I take an online master’s or get certificates?”

For this scenario, the goalpost here is actually “getting a data science job”. Naturally, if they do successfully find a data science job, then they do not need to take more online courses/certificates, or even the master’s. If you get to point B directly with what you have right now, there is no need to take a detour, the detour being those online courses.

When I get asked this question during a coffee chat, I usually suggest the following:

Personal anecdote: For some time I considered taking the OMSCS online CS master’s degree. I was already working as a data scientist at the time, and successfully delivered data science products to millions of customers. I accomplished this goalpost already with a master’s in economics. Conclusion: hold off on the master’s in CS, unless some future career move (more similar to Type 3 scenario) requires it.

[Type 2] “I want to take this Coursera course because I have little or no machine learning related experience on my resume. I really need at least something to catch the recruiter’s eye to give me a callback and phone interview.”

For this scenario, keep in mind this person (could it be you?) already has some work experience. Maybe they did project management for 3 years, or engineering for 1 year, or was teaching/researching in academia for 5 years.

The goalpost in this scenario is “getting at least an interview”. So, the easiest thing you could do here is just starting applying to jobs! It doesn’t matter if you “aren’t sure if you’re ready yet”. Don’t hold off until you’ve completed some magical number of online courses.

Do the direct action relevant to your goalpost, and if you don’t hear back after two or three weeks, then it could suggest you really don’t have enough resume line items. But you don’t know this until you’ve applied to some jobs. Don’t let your own uncertainty speak for recruiters and hiring managers. You are not them. As to what you can do if you don’t hear back from any job applications, I break down more specific details below, as even in Type 2, there are a lot of variations.

Here, I usually suggest the following, regarding taking more online courses or additional education:

[Type 3] “I’m interested in moving upward or laterally in my current data science team to a project that uses reinforcement learning (RL). How do I show my manager and coworkers that I am skilled enough in RL?”

This is great! You’re in the field, and are looking for some growth opportunities. Likely your statistics and programming skills are both at a good baseline (maybe you’re stronger at one than the other), so you don’t need to spend as much time brushing up on them as those straight out of school or transitioning from another field.

Here, I’d definitely say go for a few online courses to build your basic knowledge in the new topics, by which I mean, take online courses on reinforcement learning specifically, not general machine learning. Additionally, if you feel one of your pillars isn’t as strong as the other, for example if you’re much better at general software development, relative to the mathematical derivations of ML algorithms, consider brushing up on ML algorithms.

One thing that seems to slip people’s minds (mine included) is that the most direct course of action to the goalpost is asking your manager or coworkers if you can get moved to that new project or team! Don’t wait until you’ve taken some online courses to ask; make your interest clear now, so that they will think of you if there’s an opening on the project, or another similar opportunity.

If the above efforts are sufficient enough to get you moved to that other data science role or specialized project at work, then awesome! Your new goalpost has been achieved. If it’s not that easy, then the next big step would be to build a quick side project on that topic, using the knowledge you learned from the online course. If demonstrating your real experience in a topic can’t convince your manager / coworkers that you are a good candidate to work on that project of interest, then… online courses aren’t really the question here.

General decision making criteria

In short, here is the general decision criteria I suggest:

Conclusion - how you can determine if you should take more online courses

Now, we’ve walked through recommendations for multiple types of scenarios, and the detailed logic behind those recommendations. Of course, each person has a different educational background, work experience, and other circumstances, but I hope that with the provided logic, you can apply it to your own specific situation!

When in doubt, remember: there is decreasing marginal benefit in taking online courses. The first few will be highly beneficial, then the more you do, the less new benefit each subsequent one gives you. The same logic applies to going for additional education.

It’s easy to get stuck in a loop of passively learning, so the general idea is to take a pause if you can already hit your goalpost with the current amount of knowledge you have. Side projects or simply starting the job application process would give you a much better return on investment on your time. I hope this helps!

Related articles about "data science"