Predicting the Spread of Pandemic Influenza based on Air Traffic Data and Social Media

Abstract

Influenza is one of the most common viral infections in the world, and puts millions of people at risk every year. Currently, detecting influenza is done by hospitals reporting data, which is very slow. Modern day technology has given a rise of interest in the field of influenza awareness, especially on the topic of detection. Researchers have found that mining messages from Twitter can accurately find influenza outbreaks that correlates with hospital reported data.

Health officials need to be prepared for influenza outbreaks, so that measures can be made to minimize the impact of a potential epidemic. Rvachev and Longini has created a mathematical model for the spread of influenza by air travel, which we have implemented as a public API.

Our implementation of the model use influenza related tweets as a real-time source of influenza incidents, as well as publicly available flight data, to be able to predict spread in case of an influenza pandemic. The results created by this implementation correlates to a reasonable degree with Rvachev and Longini's original results, and have been tested on more recent data. The final system also has a simple interface, which visualizes the spread using Google Maps with an isopleth overlay.