Writing: both learning how to write, and then practicing writing. Specifically, writing as a way of working through a problem and explaining it.

I talk about this in more detail, and suggest a book to read about writing, here: https://codewithoutrules.com/2016/06/15/writing-book/

Ask yourself questions about structure, purpose and functions of parts you have spotted in anything in your daily life until you find out that your partner has a hidden "mute" button, that you just never knew existed and become excited about the value of analytical thinking.

Mathematics and puzzles/games. Take sudoku for example. It helps you to practice deductive reasoning. But the important step (I think) is when you start trying to solve harder problems and turn to math to formalize the collection of techniques you have accumulate during practice, resulting in concepts like 'pre-emptive sets'. When you transition from a collection of techniques to a theoretical foundation, you know have a much deeper understanding of the problem and how to reason about it.

You could call it 'deep learning'.

But you shouldn't ;)

Exactly the comment I was going to make.

After playing one puzzle game on commutes for a few weeks, I'll find myself almost subconsciously doing the same sort of pattern matching against bricks or tiles or the windows on a high rise. Switch to another game and it becomes a different sort of pattern matching.

The human brain, of course, is a trainable pattern matching engine. Puzzle games are great for giving it a workout and teaching it new classes of patterns. The more casual the game the better. Dots or 2048 or Mah Jong or whatever.

I find it comes naturally when I do the right activities to boost it: code, reading, discussing complex topics with friends and co-workers, etc.

Read fiction. For me, science fiction broadens perspectives and increases creativity in ways that apply directly to work.