How much mathematics and statistics?
Common issues, complaints, and questions that come up on Reddit Data Science are: How much mathematics and statistics do I need for data science, and how do I go about getting that? If I’m lacking mathematics skills would it be better to focus on the mathematics need for data science?
First let us say we believe that, as a data scientist, you need as much mathematics as you can possibly absorb at any given time.
What’s more, you will almost certainly always need to learn more mathematics and statistics throughout your career as a data scientist.
Take a course? Read a book?
Having said that, we find that common advice on Reddit Data Science is: take a course, go back to school, read a book, and this is where we disagree somewhat. Let us explain.
Some people are happy to take a course, or read a book. If that works for you then do it!
The reality is that it does not work for a lot of people, and the sad thing is that these people then feel there’s something wrong with them because they’re having difficulty coming to terms with abstract mathematics and statistics.
There’s nothing wrong with them. The fault lies in the abstracted approach of almost all college level mathematics and statistics courses. Mathematics and statistics are disciplines unto themselves, and mathematics books and instructors do not often bend to the motivational needs of beginners.
You know the drill, you’ve seen it: someone, usually a guy, often with a beard, writing squiggly marks on a board, often talking at the board, without much of an introduction as to why anyone would want to understand these squiggles, just launching right into the mathematics.
Or, you open a book and try to read it and this what you find:
Holy cryptology, Batman! Why are we interested in rectangular arrays? Can someone please tell us why we’re doing this?
As Piper Harron writes in her Ph.D. thesis: “Oh my goodness what on earth does any of this mean why can’t they just say what they mean????”
The usual mathematics and statistics courses and texts run counter to what we believe in at math4plus: that it does not have to be this way; that by experimenting with what works, with how real people really think, we can do a much better job of making mathematics enjoyable, empowering, and fulfilling. We can and must get away from the “same old methods for explanations” and discover what really works.
The Roman poet and philosopher Titus Lucretius Carus, wrote, in the first century BCE: “quod ali cibus est aliis fuat acre venenum” which translates more or less as “what is food for one may be poison to another“. This, in our experience, is especially true in the learning of mathematics and statistics.
So, if you find reading a mathematics book akin to taking poison it’s actually not your fault or your shortcoming: it’s just a mis-match between the, probably terse, writing style of the author, and your need for motivation, orientation, and understanding before getting to the technical details.
What’s the answer?
Currently your best bet is to find a mentor, someone – preferably an experienced data scientist – who can guide you through the mathematics and statistics with motivation, orientation, and understanding.
Taking a course can be good. Reading a book can be good. It depends very much on your motivation. Usually doing either with others, in a group, will help you be better motivated to continue.
At datascience.university and math4plus we believe in tailoring education to a learner’s needs. If you need help and think we might be able to help, please contact us.
Gary & Adriano