TL;DR version
This is an in-house solution that is similar to Netflix, Eharmony website which can recommend matching partners.
What do we do?
We collect user interest as variables (x-axis) and how much do they like in interest-level (y-axis) and create an interesting mapping.
By using more than one user, we can see the visual mapping between users like below:
In the graph above, you can see how much both users are common!
Our algorithm will create a score from 1–100 between these 2 users. By comparing all users, our algorithm will recommend the highest match to the users.