What it does
YouTube Grapher takes the video ID of a YouTube video and generates a force-directed graph of the comments on the video. Large circles represent parent comments and small circles represent replies. The colors are assigned based on the sentiment assigned to the comment by the Microsoft NLP API. The sentiment value of a comment is a number between 0 and 1 and positive comments have values closer to 1. Comments that have sentiment values over 0.5 appear as green circles and comments with sentiment values lower than 0.5 appear as red circles. Comments that have sentiment values close to 0 or 1 are represented by brighter circles. Circles are connected to circles that represent replies as well as the comments that came before and after them. Comments are displayed when the user hovers over a circle. YouTube grapher also generates its own Markov comments based on the ones displayed in the graph.
How we built it
Fetched comments using the YouTube Data API, parsed comments with Node JS, calculated sentiment values using Microsoft NLP API, produced graph with d3.js, produced Markov comments with Node libmarkov