I honestly don't remember how it works in Yakuza, but either way the optimal strategy is to tap repeatedly because it lets you get more tiles faster (for instance pressing D, W, A in quick succession would get 3 tiles and takes about as long as just pressing W for 1). For the post-jam update I'll see if I can add holding keys in a way that doesn't feel too slippery.
As for the music, it was definitely the toughest part because I'm not musically inclined. In the jam ver, only "beach dance" and "encounter" use a repeating pattern of notes (as in they have a note every X frames through the entire song) while the others I tried to map myself. I tried to map them such that there'd always be something audible matching the note, but it's a shame they've ended up so inconsistent and I didn't even know. For the post-jam update I'm adding more visual effects to the blue square at least so it should be clearer when the timing is, and I've added a few more frames to the GOOD and BAD windows post-note.