Love the data science take on dating apps. I wonder what you findings would be if you tried newer apps. The data I have looked at (at work), shows that the “influential desirables” are only on a dating app for the first 6–10 months of a popular dating apps lifecycle. They tend to move with each other in waves.
That means that most of those 8,000 matches are a medium to low volume of desirables. Tinder was the first real dating app, hinge was next, and bumble is the now. Bumble is on the tail end of the desirable adoption curve. I wonder what your results would be there with a different demo?