Here's some projections by a model developed by @VJL_bball:
https://laynevashro.shinyapps.io/preds/
Myles Turner projects to be a monster(18.5). Other players projecting to be stars in the league are - Stanley Johnson(10.3), Tyus Jones and Jahlil Okafor(11.0), D'Angelo Russell(11.4) and Karl Towns(11.8). From the European players Hezonja and Porzingis project best...
The regular disclaimer for small sample sizes apply...
Allow me to explain. That app actually has 2 models: the regular EWP and the HUM (Humble) model. They differ in the way they get to the final number, bur the major difference is that the Humble model takes in consideration the opinion of major mock drafters (Draftexpress and Chad Ford's big board). In some cases it will produce different numbers than the EWP model, but results should be mostly consistent. Both models compare the stats of an european of NCAA prospect to the stats similar players put up in the NBA, it's actually quite advanced and sophisticated. It also takes into account weight, height and combine athleticism values.
Here is how you interpret EWP values: EWP < 0 is a bust, 0<EWP<5 is a bench player, 5<EWP<10 is a starter and EWP > 10 is a star.
Some players the model correctly picked (as of recently): Rajon Rondo, Andre Drummond, Omer Asik, Jrue Holiday, Danny Green and Draymond Green, for instance. The model also correctly predicted international players like Dirk and the Gasol brothers, Toni Kukoc and even Manu Ginóbili.
However, the model also had some stinkers: failed to recognize Derrick Rose as a superstar (only gave him a 5.9 I believe), the EWP failed to recognize Damian Lillard although the Humble model recognized him, failed to declare Tyus Thomas as a bust (said he was a star, in fact). The model has a tendency to overrate undersized, underskilled power forwards.
As any mathmatical model, it eliminates context, so it should be used as an additional scouting tool and not the end all, be all of scouting. It doesn't take into account intangibles or the fact scoring is very, very hard to prefict going from college to pros (for instance, rebounding and shot blocking translate pretty well). It's always better than some subjetive eye test observations and crap shoot mock drafts: browse the results some more, you'll be surprised at how accurate most are.
Quick edit: the model also somewhat underrates athleticism if a prospect doesn't produce great numbers in college. That is why it underrates Rose but not John Wall.