Event Pulse

“Event Pulse” is a digital humanities research project that brings together social media data and broadcast news data in order to discover, track, visualize, and analyze the underlying forces of world events and the multiple ways they are represented through the media.  Focusing primarily on Twitter and broadcast news, the project seeks to identify the“pulse” of an event through large-scale statistical metrics, sentiment analysis, topic modeling, semantic clustering, geographic information systems, and real-time analysis. We have worked on many events, big and small, ranging from the Arab Spring and the tsunami and nuclear disaster in Japan in 2011 to the Boston Marathon bombings and the school shooting at UC Santa Barbara. The team uses a unique resource at UCLA, the NewsScape broadcast news media archive, which records and marks up more than 140 news programs worldwide, in conjunction with social media data available through Twitter’s public interface.  We are interested in how centralized and decentralized media represent events in different ways, as well as how they explain what happened and provide causal reasoning.  We also attempt to identify “inflection points” or “forks” in events, which indicate tipping points or changes in the nature of the event.  Ultimately, the research will allow any user to delve into the true complexity of an event using feeds from hundreds of broadcast news stations in conjunction with millions of data.