Analyzing social media via event facets

Abstract

Microblog is a prominent information platform for sharing experiences, discussing current events, and exchanging ideas. Many events are first reported in social media, and increasing amounts of rich-media content are associated with the posts, making them more credible and attractive. We design a rich-media analysis system to address the important challenge of sensing and exploring events from social media in real-time. The system includes a novel bilateral correspondence topic model to extract representative content and meaningful facets about events over time. It also includes a digital magazine that anchors user interactions with event facets. We demonstrate several examples from more than 4 million rich media microblogs, showing the effectiveness of key content extraction and natrual interactions with facets.