Villain has pretty typical 6 max numbers, but with a small sample size. In that small sample size he has a 10% 3-bet, so the fact he didn't 3-bet tends to eliminate AA, KK, QQ, AK and AQ from his range.
Donk betting is for donks, so the question is, "Is he a player who wants to look like a donk, or a donk?" We can't tell from the sample size. Your reraise on the flop tells
him you are serious, but he doesn't seem to care, getting 1/4 of his stack in on the term in what looks like a perperation for a shove.
He is very polarized with either 88, 33, or thinks he push you off with a flush draw or air. You are 60/40 ahead on the flush draw. and a 10/90 dog against pairs. There are 6 combos of pairs he could have and three as many flush draws (we have to assume he doesn't have AKs, AJs or 62, I'd play AT, A9, A7, A6, A5, A4, A2, KJ, KT, K9, JT, J9, T9, 97, 65.) 6 times you win .1, and 15 times you win .6 or (9+.6)/21 = 45/55 dog.
If you figure the odd called AKs preflop balances the pure air, you should shove. You built the pot to 30BBs and now shoving cost 70BB to win 100BB. Not great and definitely high variance.
Would have been better to call the donk bet denying the villain information about your hand and reevaluate on the turn, folding on the turn to a decent bet, and raising a non flush brick.