 |
Real-time Topic Exploration on Map (Lisi Chen et al.)
With the rapid development of online social media (e.g., Twitter, Facebook, etc.) and GPS-enabled devices, huge amounts of data with spatial, textual, and temporal information are being generated in an unprecedented scale. For example, Twitter, which allows users to compose tweets, has 320 million monthly active users who posted 500 million tweets per day, where 80% of the active users are on mobile. Tweets are timestamped and they can be geo-tagged by enabling the geo-tagging functionality. Tweets are regarded as one of the most important first-hand news sources and they have been analyzed for various types of human activity including bursty events, disaster management, political actions, commercial activities, crime prevention, emergency services, etc. In order to provide users with insights about real-time local or regional events, trends, and public concerns, we propose to continuously discover significant local human (SLH) activities from a steam of geo-tagged textual data. Specifically, each SLH activity is defined by a geographical region on map and a specific topic generated by an online topic model. We will investigate how to mine SLH activities in a real-time fashion by taking both mining efficiency and the quality of mined SLH activities into account.
For further information on this research topic, please contact Dr. Lisi Chen.
 |