YouTube’s Mysterious Algorithm
https://sites.temple.edu/tudsc/2019/03/26/network-analysis-on-youtube/
Ever wondered why you are watching some guy playing with slime the night before your prelim?
This article explores how recommended videos are interconnected. So, if you watch a certain video, you can be lead to a certain group of recommended videos no matter what. For example, one video can be recommended from watching any videos of a certain finite set of videos from different categories that are connected by even just one topic. The one video in the center therefore makes it possible for many more shortest paths between nodes (videos).
The YouTube algorithm has multiple massive networks that intertwine. The articles uses the name “hub” to define ego nodes. These ego nodes can be for both comment threads and recommended videos. There are so many factors that affect which videos you see personally on YouTube: likes/dislikes, subscriptions, country of posted videos, search history, watch later playlist. Every action you perform on youtube.com affects which videos will be recommended to you next. Even turning on post notifications for certain youtubers you have subscribed to will put emphasis on the content created by those youtubers. Even the creators (youtubers) have a say in what you are recommended if you click their video because they can add tags/key phrases to their videos. This is a similar algorithm for YouTube ads. Since your Google Gmail is connected to your YouTube account, I would argue that even just searching on Google when you are signed in to your Gmail will affect your YouTube recommendations.