If I wanted to define how drugs and cosmetics chemicals/ingredients overlap, could I use openFDA to help answer the drug part of that comparison? Would I be able to take the contents of all the active_ingredient fields in the drug/labeling records of the openfda beta project, parse the results, and have a list with which to work? Would it be comprehensive, timely, partially so? Perhaps naively, is there already a public list (or db) of all approved drug ingredients (not finished products) already? (Not having a 150 reputation, I could not begin a cosmetics tag.)
It seems like your idea is possible using openFDA (or dailymed data which is where openFDA gets their SPL, Structured Product Label, data, I think), but it's not going to be necessarily easy.
You can query for each and every label and extract the active_ingredients field form http://api.fda.gov/drug/label.json but it is a comma separated list and sometimes in sentence format.
Perhaps in the future this could be made into an array and you could do something like https://api.fda.gov/drug/event.json?count=patient.drug.medicinalproduct.exact on it, but that doesnt appear possible right now.
Personally, I would look into data sources other than openFDA for this use case.
