It`s possible, at least part of it for a few companies. I have a copy of Henry Hellmer`s "secret book of glass formulae" that was published in facsimile some years ago. He was Akro Agate`s first real glass chemist, worked for them (and others, like Cambridge Glass and occasionally - for special formulas, Lawrence Alley and others). Has every color he knew and every color he formulated himself, w/ notations for what they were for (marbles, tempered glass dinnerware, etc.) and the dates he first made them, the results, etc. Every one starts off with "a thousand pounds of sand" and goes from there.
I know a formula book, or at least a partial one, exists for Peltier glass, and many MFC formulas are known.
The other half would be the really hard and expensive part you talk about - pulverizing a marble to see if you could determine the chemical mix of it. But even then there`s a lot lost that`s unrecoverable - the temperature of the furnace, the drafts in the shop, even the barometric pressure can have an effect.
But I like the dream of it!