That's the Victoria 2 Culture map at the start of 1836, 392 years after the start of Eu4. Each province shows a percentage of each pop culture in said province. In the eastern bit Anatolia it's pretty heavily split between Armenians and Kurds in each province, with Turks acounting for a little less than a third in what would correspond to Erzurum in Eu4 which is the only place aside from Trebizond where there is a sizable amount in Eastern Anatolia. Greeks are also fairly spread out in Anatolia but constitute a neglible demographic past the coasts. Averaging about 5 percent in the Central Anatolia and around 16% or so around the coasts. I took a look specifically at Western Anatolia and averaged out the percent of Greeks living in each province and tried to give an average percent for the corresponding provinces in Eu4. They are as follows:
Izmit: 18% (Highest City 25%)
Hudavenigar: 13.7% (Highest City 17%)
Izmir: 14.3% (Highest City 29%)
Mentese: 5.5% (Highest City 8%)
For reference here is Edirne and Trebizond which both start of as Greek at the beginning of Eu4
Edirne: 38.2% (Highest City 56%) - This one was a bit difficult as the provinces were harder to match up
Trebizond: 22.5% (Highest City 24%)
If this at all accurate than I think it may be safe to say that the 3 provinces could start out Greek in Eu4. The percentage of Greeks in Edirne and Trebizond dropped by about 70% in 392 years. This means that the western Anatolia would have looked like this
Izmit: ~88%
Hudavingar: ~83.7%
Izmir: ~84%
Mentese: ~75.5%
That is a pretty heavy majority, even if you take into account things like people moving around/getting killed in the initial invasion. (Which i'm not entirely sure of the number of Greek peasants left dead from)
Can someone go check this out for me? Specifically how many Greek left eastern Anatolia and where they went exactly? If they all went to Greece itself that changes things but if they stopped at the Coasts (because there wasn't anywhere left to go really) that may vindicate those numbers.
Edit: I use this because Paradox had researched this themselves in (hopefully) a non-biased manner like we would.