How much Math do you really need in Data Science?
When Enough is Enough
If you’re currently studying data science, I’m sure that everything seems way more complicated than it actually needs to be. Just a quick Google search about the prerequisites of the field will uncover something that gets you anxious – and that’s math.
But how much math do data scientists use in reality? Keep reading to find out.
If you were to do a quick Google search about math in data science, you’d probably end up with a Quora post to which some math Ph.D., 180 IQ brainiac responded to in the following manner:
"Well, kiddo, you’ll need to master:
- Advanced linear algebra, Multivariate calculus, Vector calculus, String theory, General relativity, Quantum field theory, The meaning of life, Kung fu
And only then you can consider learning some basic programming and analytics."
Okay, maybe, just maybe I’ve exaggerated a bit. But you get the point. A lot of "resources" state so many prerequisites you should know in your sleep before even considering off-world abstractions of Python and machine learning.
Yeah, like that makes sense.
It doesn’t, and it pisses me off because while I deeply respect anyone with an advanced degree in math or a similar field, not all data science jobs are reserved for them.
In the following sections, I’ll break down the stuff I found most relevant when first getting into the field, in a hope they’ll also suit you well.
Learn the Intuition First
Humans aren’t meant to do advanced calculations manually, and I’m eager to know why so many universities force this approach. Like you’ll remember it two weeks after the exam.
And why do I have the gut to say this? Well, it’s because we’re in the 21st century and we have the damn computers. Those aren’t just meant for sending emails and gaming, but for heavy-duty math stuff also. It’s needless to say how much faster and errorless it is.
You, as a human, should focus on developing the intuition behind every major math topic, and knowing in which situations the topic is applicable to your data science project. Nothing more, nothing less, but this brings me to the next point.
Doing Stuff by Hand is Pointless
You’ve heard it right. Real-life isn’t school, and on your job, it’s most likely that only the end result will matter, ergo what you can produce with your current knowledge.
Your boss most likely won’t care you’ve implemented gradient descent on pen and paper first, and in Python second. And in those rare cases where he/she would care, it will probably be targeted towards you spending more time than you should on solving trivial tasks.
If you get a boner or something on writing equations, leave it for your personal time.
So, You’re Saying I Don’t Need a Math Degree?
That’s absolutely right.
Don’t get me wrong though, I think that math or math-heavy degrees are a great way to break into data science, but it’s not necessary to know everything about everything.
This comes with a gotcha, however.
If you’re at an interview for a potential data science position, and the interviewer is this guy who knows more math than you, there’s a possibility that he will give you a hard time. You should be prepared for that case. It doesn’t mean it will happen every time, but be prepared.
This has nothing to do with you not knowing what you should know (probably), it rather has to do with your interviewer being a di*khead. He’ll want to appear smart by making you look small. Ask yourself, is that an environment you want to work in?
Golden Rule
You’ve most likely heard this one before, but I feel like it needs to be repeated again and again:
Be better at programming than an average mathematician.
Be better at math than an average programmer.
It’s that simple. Data science is a mixture of fields. You shouldn’t be the best at any of them individually but instead should strive to be good enough in everything. Keep the big picture in mind.
Before you go
I just want to say one more time that this article wasn’t meant for some trash talk about math majors or Ph.D.s. As I stated earlier, I have the highest respect for them.
It’s just that when it comes to the real world, and an average data science job role, there are more important things than knowing everything about math. Math is just a tool you use to obtain needed results, and for most of the things having a good intuitive approach is enough.
Thanks for reading. Take care.
Loved the article? Become a Medium member to continue learning without limits. I’ll receive a portion of your membership fee if you use the following link, with no extra cost to you.
Share This Article
Towards Data Science is a community publication. Submit your insights to reach our global audience and earn through the TDS Author Payment Program.
Write for TDS
