A0296
Title: Nowcasting the public's response to natural hazards in the google era
Authors: Amaryllis Mavragani - University of Stirling (United Kingdom)
Konstantinos Tsagarakis - Democritus University of Thrace (Greece) [presenting]
Abstract: In the Big Data era, employing online sources to analyze and predict human behavior is significantly increasing. Internet data have the advantage of providing the revealed instead of the stated users' preferences, thus giving us access to information that would not be accessible otherwise. Online search traffic data are suggested to be valuable in forecasting, mostly in the fields of health, the environment, and economics. Towards this direction, the aim is to examine the online behavior and reaction towards natural phenomena, e.g., forest fires, floods, and tornados. Next, the correlations between online queries and the respective indices are estimated, in order to explore the possibility of nowcasting the public's behavior. Preliminary analysis northern hemisphere countries shows that the public's reaction towards such natural hazards is immediately depicted on online queries. Seasonality of the online activity follows that of the hazards, as in the case of forest fires, peaking during the summer months. The present analysis has significant policy implications, as retrieving data from online sources tackles the issues caused by the traditional data collection methods, especially in developing social indices that require real time data for immediate and robust results. Overall, in finding new, innovative ways of retrieving real time data, online query data have been shown to be of much value, as they correlate with real life data and are undoubtedly valuable in forecasting.
