This looks very interesting. As a UX Designer who was originally trained as a Classical Music composer, I'd like to comment in on a few aspects of your notational system that might be some issues and require some extensions or additions.
First of all, it looks as if this system would be more practical for relatively simple music written in keys with no more than 3 sharps or flats, with no tuplets (triplets, 5 in the time of 4, etc.) or much use of dotted rhythms, syncopations, or other rhythmic complexity, that is played at moderate to slow tempi. There doesn't seem to be any way to show how an ‘accidental’ (a note outside the current key) is an A-flat versus a G-sharp (and yes, while they sound the same on the piano, you might end up playing them differently as typically the G-sharp leads upward to the A and the A-flat leads downward to the G). You might be able to build in some extensions to handle these (an arrow showing it it’s a flat or sharp?), but I'm not sure it would maintain its simplicity and close ties to the keyboard.
Therein lies the other aspect of this system; it is so closely tied to the keyboard that it inherits some disadvantages. Mainly, there seems to be a lot of wasted space because we can only practically play 8 to 9 notes at once (in a chord) and not a ton more in close arpeggios, yet all of those un-struck keys (on a standard piano or high-end keyboard there are some 79–80 of them) appear in every bar. The current system of ledger notes, while not ideal, does have the same effect as that of a button outline that only shows when needed (i.e. one of those ‘hover’ states that is invisible except when relevant). Perhaps some way of having a sort of shorthand for those unused registers, leaving them out until they are needed, so that you could include more music without having to page-turn or scroll. Speed up the tempo or have shorter notes and you might end up with the symbols flying by (or according this scheme, up) before anyone could see them. Another keyboard/pianistic issue is the case where you’d want to indicate which hand is used — in some cases, the left hand wanders into right-hand territory (and less frequently, vice-versa). Again, perhaps this could be skipped if not necessary for most simple songs.
I realize that your goal was perhaps to provide a simpler notation for simpler music that could be read by amateurs, and that makes sense. You might find some further inspiration (and ideas worth stealing) in the shape-note notation that was written for singers of religious music in the 19th century, which had some of the same ideas, but with the goal of making a notation easier for non-musician singers (https://en.wikipedia.org/wiki/Shape_note).
I'll be interested to see how others react to this, since I'm somewhat ruined by nearly a lifetime of having to get used to the ‘traditional’ notation system, and as they say, one can get used to hanging by your thumbs with enough time spent doing it.