Never before have I been asked how much a calculation "costs," but that was a lot of this course's focus. Numerical analysis was less about learning new math and more about learning how to figure out the quickest method to solve math problems. The first half of the quarter was about speeding up matrix calculations, and the second half centered around interpolating with polynomials to approximate functions. Both were fascinating, as a lot of previously-learned math was used to prove that certain methods worked better or faster than others. I thoroughly enjoyed looking into potential applications of the class as well, since data science relies heavily on these kinds of techniques for large problems. I only wish UW was offering the second quarter of this course this year because I would love to keep diving into the subject.