If you look at the "per year" it increased again under Trump.

Why would you look at it “per year” when speaking in terms of “per president” other than to say “except trump”

Because not all presidents have served for two full terms. Examining things while ignoring the time period over which they happened does not a meaningful analysis make.

In that case, neither metric is appropriate, and we should be looking for the trend per presidential term.

O(year) ≈ O(term) / 4

(Assuming that all terms are the full 4 years long, which happens to be the case for all of the presidents being discussed in this thread)

Why would you not?