Ratings only matter to the user who chose them. They are like favorites but more. You can view content by ratings in these ways:
FAVES: Lists all content you rated decently-- that is, above average, in order of rating. Simple.
People you tend to like: Whoever you have the highest score with using the affinity formula.
FANS: People whose affinity scores for you are positive, sorted first by those whose affinity score with you is higher than any other user, with those then sorted by how high that score is. Though, this should probably be fudged a bit to cull out users who don't rate much!
FRIENDS: The idea of this category is to find people who rate things similarly to you, though I haven't settled on a firm implementation of it yet. Could be rather expensive to re-calculate your similarity with every user every time you rate something, after all. But OkCupid? has done it, so it must be possible!
ENEMIES: People who hate what you like and visa versa. This is a fun category!
SUGGESTIONS: Anything you have not rated which similar raters to you (friends category people) have rated highly.
CONNECTIONS: Users whose content you have never rated, or not rated much of, whom your friends have rated highly.
Affinity formula: Every time a user (A) rates another user (B)'s content it recalculates A's affinity for B's work. The affinity formula is intentionally biased toward more faves. It simply adds up all of A's ratings on B's content linearly, although offset so that negative ratings are actually bad. For example, if you are rating 0 to 5 meaning 0-loath 1-dislike 2-no-opinion 3-enjoy 4-love 5-worship, you would subtract 2 points for each one so that no-opinion does nothing and negative ratings are bad.
Google Maps integration: All of these sections that list users should provide a map full of pushpins indicating any of these users who have provided their country name and zipcode.
