That is the solution in essence, but it is hard to read, not to mention that debugging is almost impossible as is.
Python is a very verbose and explicit language. We shouldn't neglect that in favor of shorter code, I think.

Also, it "kills 2 birds with one stone", because it checks if arr is None as well. It could be subtle, but if one gets used to it it really becomes an elegant way to check for empty list and None.

Because you want the count of the positive numbers. Each x that is positive gets represented by 1 in the new array. A sum of the array with 1s is essentially a count of the positive numbers.

Doesn't work for None

Isn't this running all operations rather unnecessary?