Linear Algebra for Data Science
kyunghyuncho.meIf I were to teach linear algebra I would stick to the 3D graphics approach. Maybe as far as including labs for implementing a ray-tracer.
It is just the most fun, intuitive and eye opening application of basic linear algebra.
Don't forget the fun!
Data science applications can come later.
Do you have any favorite resources (can be books, courses, blog posts) that teach with this approach from? I have been diving more into 3D and the shortcomings in my mathematical background are starting to show.
Good question. I learned it the hard way long ago. But it was still the 3D space in mind, although not with computers, but hand drawn pictures.
I saw my neighboring programs with supposedly simpler linear algebra courses struggle with gaussian elimination and thought it was utterly boring. My course was probably also boring but at least the 3D space made it come to life. It gave a meaning to translations and matrix multiplications.
I think this course looks promising:
http://immersivemath.com/ila/index.html
The times are so much better now, should be so much easier to learn these things with all the resources available. (Although I realize that is an easy thing to say.)
Maybe these two submissions can help you dive further into the topic:
https://news.ycombinator.com/item?id=19264048
https://news.ycombinator.com/item?id=40329388
Seems like this guy gets a lot of praise but I have not downloaded the book (I should probably stop this rabbit-hole and eat breakfast and let you to it).
I didn't go to university for CS. Instead I did physics. Because of that, learning/remind myself of this stuff was relatively easy, and so I did it.
And let me tell you, it didn't actually made me a better data scientist/model builder, for the same reason that learning how to implement some tree traversal didnt make me a better programmer.
I think I'd argue with both points actually. Though it also depends on the nature of work that you do. I imagine good intuition with linear algebra is crucial for novel work. If you're implementing a model off an arxiv paper maybe not so much.
Also remember that you (GP) already acquired that intuition in your youth. Most things you learn look trivial in retrospect. It may be that you already had some use of that.
And if you had not learned those things you may have started worrying what this magic thing you don't know so much about is. That can also hinder performance.
(Latd edit since I wanted to add). Another reason it is good to know stuff is to be able to dispel co-workers trying to bullshit. They are everywhere and I don't think they are trying to be mean, it is mostly insecurity. But particularly in this field I think people will start throwing buzzwords around to confuse and lead you away from what is simple and important.
But GPs comment is fine as a thought provoker.
This seems to be a more hands-on linear algebra intro, starting with matrices and building up from there. Note I've not actually read the whole thing, just skimmed it.
Something similar worth looking at is: https://ocw.mit.edu/courses/18-065-matrix-methods-in-data-an...