MOOCs won’t get you a job in data science. I know, it’s frustrating, because everyone is looking for the magic pill. It would be fantastic if we could just watch a couple of hours of video, write a couple lines of code, and voila.
It’s doesn’t work like that in the real world.
For some reasons that I will explain in this article, MOOCs doesn’t allow the participants to transform the knowledge they learn into a concrete skill. If you’re already raging in front of your screen convinced that an employer will pay you 100K/year because you got that 9.99 certificate of completion, then I suggest you close this window because it won’t get better.
MOOCs dont build skills
They simply don’t. MOOCs are good for sharing knowledge , but that won’t help in a real work environment.
MOOCs are most beneficial to intermediate and senior professionals who want to get knowledge quickly. They can take that knowledge and practice what they learn by applying it on real world problems. They can then get feedback from colleagues, and that’s how they transform their knowledge into a skill because of their experience and environment.
Lesson #1: MOOCs give you knowledge not skills.
There’s a major difference in between knowledge and skill. For example, knowledge is you being aware that you can use R to fit a probabilistic distribution to your data. Skills is you opening R, throwing the data in, performing the fit, and explaining me in details what’s going on and the options to calculate a 95% interval if the data is not normal. Also, you should be able to do that with clean data, messy data, special cases, with or without multicollinearity, with or without regularization, etc. You become skilled and employable only at that level.
The real way to learn is to put what you memorized into practice and iterate by getting feedback from colleagues and mentors. Only then will you reach the level needed in the Data Science world.
The problem with MOOCs is that you don’t have the chance to go through that iteration process. Discussing the concepts with other students in forums is really not enough.
Keep in mind that skill is 20% knowledge and 80% experience/intuition. With MOOCs you’re only getting the first 20%.
Value comes at a price
Yes, the law of supply and demand still applies in the Internet age: you get what you pay for.
It’s easy to fall into the trap and think you can take 9.99 MOOCs and then get 100k/year jobs. But i want you to reflect on the true value of MOOCs. Is the value really in the videos and the assignments? No.
The better way to learn would be that the instructor review your work and give you feedback, using his own personal time. The real value in the instructor is in his time and skills, not in the videos and assignments. The problem is that you won’t get access to either his time or his skills, because you paid 9.99 and he’s not running a charity.
Lesson #2: You get what you pay for.
If you want quality and build real skills, you need to pay for it. For example, in my SQL course I review the projects of participants myself and communicate with them after each assignment to make sure they build a skill.
Guess what though? I’m not selling the course 9.99, because you’re getting real value that can have a big impact in your career.
Statistics don’t lie
According to a report on Johns Hopkins University’s data science specialization offered through Coursera, there were 1.76 million course signups in less than a year after the specialization launched in April of 2014. Of the 1.76 million course signups, only 71,589 Signature Track verified certificates were awarded.
Within that 4%, what’s the percentage of people for who the specialization really made the difference in between unemployment and employment? I bet that percentage is very small.
Also, the grades for MOOCs are just unreal. In a typical university class you might have 10-20% of students who end up with As. Look at that:
That’s for the edX 6.002x: Circuits and Electronics (Spring 2012) program. That’s a clear indication that the course is too easy. And again, the percentage of registrants who got their certificate is… you guessed it, 4%.
I could go on and on like this but i think we can conclude that:
- MOOCs are not engaging enough, because more than 95% of people don’t complete the course.
- MOOCs lack in depth (next point) as most registrants get As which is abnormal.
MOOCs lack in depth
Remember the last time you watched your favorite sport and thought “this looks easy”? You then went to play it with confidence but realized you were wrong.
The same thing will happen when you watch videos in a Data Science MOOC. You will see the instructor do a regression and think that it’s so easy. After all, you just have to write 3 lines of code, right?
Then, you will get in an interview or at work, have a real dataset in front of you and be lost. Maybe the data is not clean like you’re used to, maybe there’s seasonality in the data and you don’t remember how to deal with it.
The problem is that MOOCs always stay high level. They don’t have the time nor the right medium to teach complex and rigorous concepts. That requires questions, challenge of ideas, a discussion basically.
Lesson #3: MOOCs stay high level and don’t contribute to build concrete skills, especially for beginners.
An alternative to MOOCS
MOOCs will simply not be of help for you to get a job in Data Science. Employers don’t look at MOOCs on resumes, and it can even make you look bad.
If you don’t already have a job in data science I highly recommend that you consider investing in yourself. A good idea could be to either pay for formal education, or to register in courses like mine that gives you access to a tutor.