When are matrices a useful data structure?

It seems like it is almost always easier to work with a data.frame/tibble than a matrice.

If the matrix is not super big, it's easier just to pivot it longer, and then do whatever manipulation that you need with tidyverse functions, and then convert it back to a matrix if you need the output in a tiff format.

What do other people think? What are the situations where matrices are a more useful structure?

Can you provide any examples where it is easier to convert a matrix to a data frame, do something to it, then convert it back to a matrix?

If you struggle with the syntax of applying functions to matrices, then I can understand that you might adopt this method, but matrix calculations should be more efficient and concise than converting back and forth.