Multifaceted Event Analysis on Cross-Media Network Data

People are discussing about events such as shootings, protests, flight crashes and new public policies across news media and social media. Different media sources reflect different facets of the events. In this positioning paper, we point out the importance of integrating various data from the multiple media sources for understanding/analyzing the events. Here the big data’s “Variety” issue can be resolved by extending the information network representation to a" cross-media information network" but how to generate comprehensive event analysis from the network’s components? We propose a multifaceted analysis framework that models four critical facets to fully understand the events across social media network and news media network. The facets include media type for differentiating event information sources; content for discovering events from unstructured text; sentiment on quantifying user reflecting on events; and, time for turning events into dynamic evolving objects. We performed our framework on a real data set from Twitter and Google News, which creates interesting and useful event description and visualization for human inspection.