Regardless of what platform you're on, you want your emoji to be similar enough to those of other platforms so that intent and meaning are transmitted accurately. This is one of the problems with emoji overall - they specify very approximately what visual each keycode should correspond to, which means that you could have two valid implementations that result in very confusing / inaccurate communication between users on each platform, since they're sending one thing and the other side is seeing something quite different.

That is not a problem unique to Emojis though. Some people construct multi-symbol things that would make a screen-reader call them what they are instead of what they "mean". Take for example the shrugging man made up of slashes, with a Japanese hiragana right in the middle of it!

It is a fundamentally wrong assumption that Unicode made: they thought if a code existed for everything, people would always use the code they are supposed to use. In reality, people seem more likely to choose symbols based on what each symbol looks like, and make their own composed "words". You practically need machine learning to correctly decipher what each string of symbols really "means".