Densimeter correction: Alcohol percentage

Online you find very poor informations about densimeter correction depending on alcohol content in beer, just tons of formulas about apparent vs. real attenuation with Plato and Brix scale, that are almost worthless.
If my saison FG is 997, I know it's very dry, but wich part is the alcohol, and how many residual sugars do I have in the beer?

Better answer would be: start using a rifractometer.
But I did this intead: I measured density at different dilution of cheap Vodka (37,75% Alc.), more specifically 18,87%, 9,43%, 4,72/, 2,36%, 1,18% alcohol. Then I plotted the table here, obtaining the following curve.

Here is the final table, with Alc.%, density, and the points to be added to densimeter

Qui la tabella, con la percentuale di alcool, la relativa densità in acqua, ed i punti che bisognerebbe correggere quando si legge la densità finale dal densimetro. 

Alc %	density	correction
1	998,155	1,84
2	997,11	2,89
3	996,065	3,93
4	995,02	4,98
5	993,975	6,02
6	992,93	7,07
7	991,885	8,12
8	990,84	9,16
9	989,795	10,20
10	988,75	11,25

So, my barley Wine OG 1094 that after 5 days was at a measured density of 1024, is not at 9% Alc., but if I correct according to the table it, is around 8%, and the apparent density of 1024 correspond to 1034: that means 90g/litro of residual sugars. So the fermentation is not finished and it will need some more days to get the real FG...