Monitoring social network user density variations in areas of interest

Mapping the population’s density is an important challenge for crisis management, specifically, early warning systems. Its assessment is key to determine the risk of an event and plan the emergency response in crisis situations. One shortcoming of current systems is that the population density is determined statically, mostly based on census data. We propose to dynamically assess the population density using data extracted from social networks with geo-referenced information, such as Twitter and Instagram. Not only do users constantly update their status on social networks, but a growing percentage of the information that is shared is geo-referenced, or can be analysed and related with a geographic location. The methodology we present in this paper will enable a dynamic map of social network user density variation in a given geographical area. We present the methodology, and two preliminary studies that assess the quality of the information on Instagram, and conclude that it enables the detection of a variation in the user density on two different scenarios.