Facebook Friend Suggestions
https://www.theguardian.com/technology/2016/jun/29/how-does-facebook-suggest-potential-friends-not-location-data-not-now
This article talks about the way Facebook suggest potential friends to appear on one’s Facebook account. More specifically it looks about the types of elements that the suggestion algorithm uses to make these suggestions. In the past, in addition to simply offering suggestions based on who has a significant amount of mutual friends, aka making simple triadic closures, the algorithm previously took past location information to make suggestion.
While Facebook claims now to have stopped using location, due to an outcry of privacy concerns, the idea brings up an interesting discussion on triadic closure as well as weak ties. For example, a person may have two distinct networks, say high school friends from their hometown and co-workers in a new city. Thus, while that person may be very embedded in one group currently and their may be many strong ties and triadic closures, because of this location variable, they may still get suggestions with whom they share few neighbors. This creates the opportunity for many local bridges to be created in networks and ultimately the multitude of local bridges will create a network of regular bridges with smaller spans. Highlighting this phenomenon, the article also references a study internally conducted at Facebook that concluded the average degrees of separation was just 3.57 amongst it’s over 1 billion users.