The definition of “safeties” varies. Some people here like to think of them as true sure things, thus the automatic admission criteria, understanding their FA in advance, etc. That’s perfectly fine, but many others use safety much more broadly to mean simply extremely likely. If you are going to make safety 100% guaranteed, you almost need a fourth category for the extremely likely that are different than the targets. None of my S’s friends nor his school’s counselor interpreted “safety” as guaranteed. What they did do is make sure to apply to more than one safety just in case they didn’t get accepted to one. I didn’t hear about a single kid at my S’s school for whom this didn’t work out and they didn’t get in to any of their safeties. Of course you do hear about people nationally who don’t get into any schools but when you look at their safeties they usually were targets to reaches by any objective measure.
Anyway, all that said, from your original list I would call Case a safety and USC a target based on your stats and reviewing the typical GPA and score data. Case was one of my S’s safeties (added at the last minute when they waived the application fee and required no supplements) and he was accepted and offered a very generous merit scholarship. His GPA/scores were not better than yours. Some people will say no school that accepts only 30-some percent can be a safety. All of my S’s were and he got into all of them and was offered merit aid at all that offer it. He applied to 3 just in case.
The Common Data Set is awesome. Many interesting things in it. I will caution though that schools are elective in filling it out and some don’t due so comprehensively. So their data is not always complete, including how many graduates per major. We encountered this pretty frequently where if you added up all the degree stats they were a fraction of the graduation rate.