Another option is to just recreate the whole model, the data and the objects. You would probably learn more doing it this way and expose anything that is hard to see within the data area. We recently did this and we find some things with there were calculations set for unknown reasons that were exposed. Also, part of the modeling is considering what/how to transform your data and apply them to master data sets.
But a snapshot would work. We do this a lot for "what if" model changes. We then bring the information to our steering committee for approval before we make the model change in production.
Good luck!