Yes - recursions are fascinating, however, be careful with them. Recursions could be double-edged swords...
Consider that each recursion level costs in memory on the stack. So if you have too many levels (self invocation) you would end up with a crash on lack of memory (or slowing down your whole computer as your code exhausts whole of the available memory).
In this specific case it is probably fine - we assume a number with a reasonable number of digits.

WoW

Oh neat!

We learned new stuff.

First and foremost "return n if n < 10" is the base case.

The map function applies a given function (in this case "int") to every item of something that is iterable. -> new for me!

Using the string function on the input n, you turn n into a string which is iterable. Therefore the map functiion can now apply a given function to each character of a string. -> the clever part!

Once the map function is done applying int to every digit, it is then all summed together.

WHAT

I mainly meant efficient in the code golf sense of not requiring many characters. :) Yes, the floor division method is more performant.

Efficient way to get the digits. I'll have to remember that one.

Oh that's clever though! You figured out a way to get it done!

yep)

But that's being super-pedantic since you're returning a list and not an array. It's a shared description. Imo returning a tuple in this context makes more sense for Python.

Even though they say return array not tuple test cases are passed.
Here you return tuple instead of array