From a Self-Taught Data Scientist, Here's Why You Shouldn't Use Bootcamps.

Opinion

As a preface, not all bootcamps are bad. And learning with bootcamps is better than nothing, which I applaud! My goal is simply to share my opinion about better alternatives that are out there.

Introduction

I’ve been in your shoes before.

Without any sort of STEM education and without any sort of guidance, I had to find my own resources and carve my own path to learn Data Science. It was certainly difficult getting started with such little direction.

In my previous articles, "How I’d Learn Data Science if I Could Start Over (2 years in)" and "5 Things I Wish I Knew When I Started Learning Data Science," I shared some of my biggest insights and reflections that I gathered throughout my journey.

But one thing that I didn’t talk about in my previous articles that I wanted to talk about today is bootcamps. I’m not a fan of bootcamps and I generally don’t recommend bootcamps to learn data science. Like everything else in life, there are a couple of exceptions which I’ll talk about later, but using bootcamps to learn data science definitely set my back in my learnings and in my career.

Here are my reasons why.

1. Bootcamps are a sub-optimal solution.

A strategy that I constantly use in my Education, in my career, and in my life is asking myself "what is the objective of doing this?" And so, I found myself asking "what is the objective of learning data science?"

Personally, there are two objectives for myself:

To develop and refine my knowledge in data science
To gain credibility to show employers that I’m are a qualified candidate.

Now with these two objectives in mind, consider the image below:

I’m not going to go through each point, but you can see that there are a number of alternatives that have a much higher impact in achieving the two objectives that I outlined above.

You might be thinking "well aren’t bootcamps good because I’m learning AND getting a certification?" and the answer to that is no. Generally, bootcamps are extremely surface-level, explaining only the "what" and not the "why". And in terms of gaining credibility, bootcamp credentials get you nowhere – it’s what you do with it that matters (see point #3).

And to tell you the honest truth, I made the same mistake when I started out. Initially, I cared so much about "looking" certified on paper that I was more focused on piling certifications and less focused on developing my knowledge. What I got out of that was wasted time and useless certifications.

My goal isn’t to try to shut you down from learning. I understand that everyone has their own learning methods that work best for them, which is great. As I said in the preface, learning with bootcamps is certainly better than not learning at all! All that I want to do is make sure that you’re aware of more effective alternatives that will help you achieve your goals much faster (aka. becoming a data scientist).

So if you feel like you’re chasing certifications and not education, I strongly recommend that you take a step back and ask yourself what your objective(s) is (are).

2. Bootcamps are misleading.

If a bootcamp ever advertises that "you can be a data scientist in 10 weeks," shame on them. This is a completely misleading statement that insinuates you will have learned almost everything that you need to know after a 10-week bootcamp which I simply don’t agree with.

Learning data science is a life-long journey – it’s never-ending. A well-rounded data scientist requires knowledge in several fields, including programming, statistics, mathematics, and domain knowledge. **** I know senior data scientists and VPs of data science that are still learning every day despite having years and years of experience. And so, if you’re not the type who’s willing to learning continuously for the rest of your career, well, I’ll leave the rest up to you.

Bootcamps are also misleading in terms of the level of knowledge that you learn. I’m generalizing, but most bootcamps teach the bare minimum that you need to know for a given topic/subject. Completing a bootcamp does not make you an expert in a subject, it has simply opened the door for you to explore in more depth.

This goes back to my first point, but if you really want to learn a concept inside-out, I strongly recommend other resources, like YouTube, free university courses, or books.

3. It’s not certifications that get you a job, it’s what you do with it.

Related to my second point, don’t be misled that a certification from a 20-hour bootcamp will get you a job. I would even argue that a master’s degree in data science is useless if you have no credibility to support it.

Looking at my diagram again, there are three ways to gaining credibility aside from obtaining a Master’s or a PhD:

Kaggle competitions: In my opinion, there’s no better way of showing that you’re ready for a data science job than to showcase your code through competitions. Kaggle hosts a variety of competitions which involves building a model to optimize a certain metric.
Personal projects: There’s no reason to have to wait for a job to get experience when you can complete personal projects on your own! If you need some inspiration, check out my list of data science projects.
Open-source contributions: Similarly, this is a great way to work on a larger project that’s already established. It also shows that you can use version control, work with a team, and collaborate.

In summary, bootcamps aren’t enough to showcase your credibility.

How should you get started then?

I’m not going to tell you how "you should" get started. Rather, I’ll tell you how "I would" get started.

Start with statistics. Of the three building blocks, I think statistics is the most important. And if you dread statistics, data science probably isn’t for you. I’d check out Georgia Tech’s course called Statistical Methods, or Khan Academy’s video series. StatQuest also has an amazing inventory of videos which I highly recommend that you check out.
Learn Python and SQL. If you’re more of an R kind of guy, go for it. I’ve personally never worked with R so I have no opinion on it. The better you are at Python and SQL, the easier your life will be when it comes to data collection, manipulation, and implementation. I would also be familiar with Python libraries like Pandas, NumPy, and Scikit-learn. I also recommend that you learn about binary trees, as it serves as the basis for many advanced Machine Learning algorithms like XGBoost. To learn SQL, I recommend Mode’s course which is quite extensive and well done in my opinion. As for Python, I recommend learning by doing, as in completing a small project of some
Learn linear algebra fundamentals. Linear algebra becomes extremely important when you work with anything related to matrices. This is common in recommendation systems and deep learning applications. If these sound like things that you’ll want to learn about in the future, don’t skip this step. Khan Academy’s course should be more than sufficient.
Learn data manipulation. This makes up at least 50% of a data scientist’s job. More specifically, learn more about feature engineering, exploratory data analysis, and data preparation.

Thanks for Reading!

I want to reiterate that I am all for learning, and I appreciate all types of learning styles that people may have. I only shared this article to bring awareness regarding possible alternatives that might help you achieve your goals faster. If you don’t agree with me, that’s okay! Agree to disagree 🙂