Spatial And Temporal Analysis Of Twitter: A Tale Of Two Countries
02 October 2014
People tend to share information with their peers in order to inform them about their current situation typically by using social media services(e.g. sharing their latest news over Facebook or Twitter). This has become a huge part of our social life. During crises this behaviour is even greater because it allows people to reassure their peers (followers and friends) of their well-being in a quick and easy manner. Lately social media services have been also used for another purpose during crises, that of informing oneself over the current evolution of the crises. However obtaining relevant information over social media can be a difficult challenge as anyone can post any information. Filtering the ow of messages such that only relevant information are remaining is critical in times of crises. We propose a spatial-temporal model that collects the data from Twitter. The data is then further processed to evaluate the density of tweets surrounding the area. Furthermore we evaluate the possibility of shared user accounts by determining the physical distance and velocity between messages originating from the same user account.
