# How can I effectively learn and master Math and Statistics for Data Science?

I completed a BSc in Computer Science recently and am going on to do an MSc in Data Science. However, the only focussed math module I had during CS was in the first year and I didn't do too well. I really want to ace it this year so I'm trying to get spend some time getting it. I've got a 'Content outline' which I've retrieved from the UK University's website so I have a list of things to do. Issue is, I know I'm gonna struggle and more sure that I'll forget what I learned if I ever learn it. How can I concrete the newly found knowledge in my brain. Is there a (fun) or repetitive method that anyone here have had success with when learning math?

Statistics is new territory to me too (I think). Any suggestions, recommendations or advice is greatly appreciated.

Thanks

• Try to answer some questions on the stack sites. If you're doing it wrong, someone will take joy in telling you. Aug 23, 2021 at 3:58

You could try writing careful and detailed notes for yourself, and by "careful and detailed", I mean something you think even your former teachers would think is good and something you would feel comfortable in lending to someone else who needed to review the material. Maybe use two or three different colors of pens, if the notes are handwritten -- one color for problem statements, another color for problem solutions, another color for facts/theorems to remember, etc.

The reason I am emphasizing the quality of the notes is that this will force you to spend time thinking carefully about the material, and thus you will likely retain it better. Also, the notes will be something you can refer back to if "I'll forget what I learned if I ever learn it" comes to pass, but my guess is that the mental calm you'll get from preparing the notes will reduce your fears, thus allowing your memory to function better.

Also, depending on the subject matter and how well you know it, you could write problems on index cards, with solutions on the reverse side. This would work well, I think, for probability and combinatorics problems, since most of that involves practice in identifying how to solve the problems, and not with learning a lot of theory stuff. The index cards can then be ordered or grouped in various ways, in case your initial way of ordering/grouping them changes as you become more familiar with the material, something that is difficult to do with pages and pages of ordinary handwritten notes (because you'll have to decide in advance whether topic A comes before topic B, etc. when writing). Of course, for digitally-typed notes, you can rearrange sections and problems even after writing them.

By the way, nothing about this is specific to statistics or data science. I used the notecards method in Fall 2011 when I spent a few months studying Hilbert style propositional logic systems (e.g. see this answer), and I wanted to keep track of those results that could be proved using only modus ponens and the deduction theorem (e.g. the positive implication fragment of intuitionistic propositional logic), and then those results one could additionally prove in the slightly larger system of Johansson's minimal negation implication logic, and then those results one could additionally prove in intuitionistic implication logic, and finally those results one could additionally prove in ordinary implication logic. Regarding written notes and such for myself on various topics, I've been doing this since middle school (ages 12-14, 1971-73).

First, I would try to emphasize accessible texts, exercises, etc. If you are weak, you don't load 225 on the bar and start benching. You absolutly SHOULD get in the gym. And will have faster gains than someone who is already very strong. But you need to be progressive. Going too heavy will hurt/discourage you. Similarly with studying something you're weak in. Be progressive (not politically, but "bit by bit"). In particular, be wary of many people on MSE who are sooper smart in math and weak on training weak students, who prescribe books/courses that are too tough and/or theoretical for weak/applied students.

Second, take a look at the course descriptions, syllabi, and texts and see what math is required. Emphasize the first semester. See where you think you might have issues/not.

Try to stay on top of the courses overall. Pre-read the lessons in the book (actually do this) and (try to) work the problems before the lectures. Don't get behind, don't cram. Go to office hours. Etc. A stronger student might be able to survive with a non-optimal study regime, but you are already fighting with one arm behind your back.

Write down your steps during any derivations or homeworks. If you miss something because of a "stupid mistake" rework the problem all over again and do it correctly.

I would not despair on the math. Realistically stats is a blizzard of formulas. But at the end of the day, if you have intuitions about some basic concepts, you can do most of it (for example least squares regression is "like drawing a line through the data").

Hang in there and have a little fun. Hard work pays off. Really, it does. You can do this.