First of all (you already notice, this is going to be shit-sandwich-feedback): I really like that you just went straight for a proof-of-concept to show to people how cool it would be to be able to identify events (e.g. conferences) by persistent identifiers (e.g. DOIs). I really like the idea, and agree that it is a sensible thing to do. One can also use DOIs, I don’t see problems there.
I do see a problem, though, because I feel that you “creatively re-interpret” (which is: misinterpret or misuse, in harsher words) the DataCite data schema. Why is that a problem? Because people rely on the schema, and they want to have reliable information just where they would expect it.
You are right in that there is an “Event” type, and you used it to describe an event, i.e. a conference. But to the best of my knowledge, this is not what DataCite thinks of an “Event” – e.g. if you look here: DataCite Event Data The ultimate reference is of course the Documentation: https://schema.datacite.org/meta/kernel-4.4/doc/DataCite-MetadataKernel_v4.4.pdf
But: I believe that DataCite would (or: might?) be open to discussion about this. The extension of DOIs and/or PIDs to all kind of things (events, but also scientific instruments, grants, you-name-it …) is very popular right now, and the logical next step. I believe you are at an interesting point in time with your idea, just at the right spot.
Let me know if / how I can support you in bringing the idea to mint PIDs for events to life!