Facebook and Information Cascades
https://www.technologyreview.com/2014/03/24/173582/the-curious-nature-of-sharing-cascades-on-facebook/
Even though Facebook’s popularity has reduced over the years, it is still one of the most used and famous social media sites in the world. Facebook was the main player to change the social media. The features it provided since the beginning were revolutionary and transformed the way people interacted with each other. Posting pictures, uploading videos was never this easy. In this blog post, I will use a topic which was taught to us in class- information cascades and hows its used in facebook. Information cascades exist when people make decisions depending on other people’s decisions.
Information cascades on facebook determine which picture or news is going to go viral or not. The post usually gets viral when it gets a large number of likes and generates a good amount of engagement. Some content gets more attention than others which leads to cascades in which the number of re-shares can be a lot. It is very tough to predict whether a post would become viral or not. Various teams have claimed to have found ways in predicting the popularity of the post right after it is posted. But it is still tough to predict the popularity with conviction. Now, due to the work of Cheng at Stanford University, we have been able to understand that it is very difficult to predict popularity of a post based on conventional methods.
Predicting the popularity of a post has various factors which are difficult to determine. But, Cheng found a way to determine the popularity of a post with accuracy and can be used to form judgments as well. Cheng’s team analysed the way photos were shared on facebook for over 28 days, and they ended up looking at 150,000 posts. They made a node for every person who shared a particular image, and formed an entire network. They used machine learning algorithm to search for the features of cascade and predict the outcome.