Sophy: a morphological framework for structuring geo-referenced social media

Social networks have played a crucial role of information channels for understanding our daily lives beyond communication tools. In particular, their coupling with geographic location has boosted the worth of social media to detect, track, and predicate dynamic events and situations in the real world. While the amounts of geo-tagged social media are apparently increasing at every moment, we have few framework to handle spatiotemporal changes and analyze their relationships. In this paper, we propose a framework to understand dynamic social phenomena from the mountains of fragmented, noisy data flooding social media. First, we design a data model to describe morphological features of the populations of geo-location of social media and define a set of relationships by using differential measurements in spatial, temporal, and semantic dimensions. Then, we describe our real-time framework to extract morphometric features from streaming tweets, create the topological relationships, and store all features into a graph-based database. In the experiments, we show case studies related to two typhoons (Neoguri and Halong) and a landslide disaster (Hiroshima) with real tweet-sets in a visualization way.