This project aims to show a different picture of the data behind the news, looking at how we analyze, represent, and interact with it. Video content is constantly created and added to the public archives, but there is never time to watch it all. News Graph explores a new method for interacting with news media. By analyzing the words that are said, extracting entities that appear, and finding the connections between them, we are able to map connections between video segments. Each connection represents two entities that were mentioned in the same video segment, and a video segment can be mapped to a number of connections.