The Language of Likes in Facebook



This project aims to build a friend recommender system for facebook, by analysing the 'likes' of a facebook user profile. The data for the project was aggregated using the Facebook Graph API

The graph data structure was developed in python using the 'networkx' package. All visualisations were developed using Gephi.

Below is a visualisation of the likes of four facebook users. Based on visualisations generated like the one below, one can draw inferences, on user interests, and how they link people together on facebook

Another module of the project involved the analysation of how the liked pages, scale in a 'like network'. Below, you find a tree like structure, rooted at a facebook user, as we delve into higher levels of recursion, we find an exponential increase in the number of pages. This particular visualisation maps pages upto three levels of recursion and contains over 200K nodes and 400k edges.