A Spatio-Temporal Approach to the Discovery of Online Social Trends

Abstract

Abstract. Online social networks (OSNs) have become popular platforms for people to interact with each other in the cyber space. Users use OSNs to talk about their daily activities, mood, health status, sports events, travel experiences, political campaigns, entertainment events, and commercial products, among other things. Conversations between users on an OSN site could reflect the current social trends that are of great interest and importance for individuals, businesses, and government agencies alike. In this paper we design and develop a comprehensive system to collect, store, query, and analyze OSN data for effective discovery of online social trends. Our system consists of three parts: (1) an OSN data collection engine; (2) a spatio-temporal database for storing, indexing, and querying data; and (3) a set of analytical tools for online social trend discovery. We demonstrate the effectiveness of our system using a recent result of predicting seasonal flu trends using Twitter data.