I remember the moment I realized I needed to learn how to play the piano.
[Psssst. Keep reading; this is still about maps!]
During my second semester at Berklee, I found myself with a bunch of fellow art-nerds, sitting in an apartment watching a video [VHS whaaaaa!] about a dance piece collaboration between Garth Fagan and Wynton Marsalis called "Griot New York". Wynton rolls into a rehearsal toting his horn in a gig bag and rolling a full-size yamaha keyboard behind him. As the voice-over goes into dance/music duality overanalysis mode, I see that Wynton is playing the trumpet with his right hand and laying down chord progressions on the keys with his left hand.
Sure, I'd heard about Dizzy Gillespie, John Coltrane, Charlie Haden et al. doing their composing on piano. I'd gotten all sorts of pressure to put down my horn and practice the keys regularly as a way of understanding the harmonic roots of the music I wanted to play. But this was live proof - in the form of one of the greatest virtuosos of a generation - that a melody instrument could only take you so far in both performance and composition before you needed Western music's chosen tool for getting at the harmony underneath.
The piano is a compiled language, to be sure - several steps abstracted from the wonder that is our perception of tone and harmony. But it fits the question I've heard time and time again from GIS analysts, and seen in blog-nonsense form hither and yon:
Do I really need to learn python/piano?
Yes. Your advancement as a scientist/artist will not be possible without it.
Let me close with another question brought on by this crude metaphor: If the piano is Fortran, does music have a kernel? :)