notes.husk.org/likes images.

Project FeedVis 

Feedvis is a research tool developed by Karrie G. Karahalios and her team at MIT to expose to Facebook users how their newsfeed behaves. They built it with the Facebook API and it compares the posts a user sees on their newsfeed side-by-side ALL the posts that their contacts have made. The study revealed that 25 out of the 40 participants were not aware that the newsfeed algorithm even existed. Some of these participants were upset to discover that they were missing out on posts from their family, and many had assumed that a lack of posts meant that their contact had stopped posting. You can watch here or read here for more details.

Users were astonished by how long the left column was and how many things are hidden from Facebook. For most subjects, this was the first time they were aware of the existence of an algorithm… 

Over the course of the study and visual narrative, participants were initially very shocked and angry. Over time, they started reflecting on the nature of the algorithms, creating theories for how the feed might have filtered their information.

After learning about the algorithm, when users went back to Facebook, they reported using Facebook’s features more, switching more often between top stories and most recent stories features, being more circumspect about their likes, and even dropping friends.