Zooming in and Out Our Society: Discovering Macro/Micro Events from Social Media

Abstract

Social event detection is regarded as an useful tool for understanding our society and more importantly for providing people with various smart city services. However, given a large number of social big data, it is hard to find out the meaningful patterns with traditional statistical analysis. The aim of this concept paper is to present how to generate spatiotemporal constraints for analyzing social data. Particularly, To discover the hidden social event, spatio-temporal constraint generation process consists of three methods, which are i) generating spatio-temporal focus (STF), ii) allocating social texts to the STF, and iii) discovering social events in each STF. The STF can be composed by clustering social texts which are topically and spatially associated. To make the STF dynamic and incremental, we use windows for allocating a social text to an adequate STF. Lastly, the hidden social events are discovered from the STF on the basis of keywords and temporal distribution of the social texts. Although, in this study, we limited the proposed method into analyzing social media, it could be extended to discovering events among all possible entities (e.g., people, sensors, and so on).