This project is dedicated to building an Early Warning System (EWS) for avian influenza in Canada. Multiple data sources including Google trends, Google news, number of Reddit posts, number of Facebook posts, air qulity, and weather parameters are combined and used in deep-learning models to accurately forecast avian influenza outbreaks by up to 14 days in advance. The machine learning pipeline which includes collecting data in real-time, training and testing models, forecasting outbreaks, and visualizing the results is automatedd for both country- and regional-level predictoins.
For detailed information on the datasets and methodology, please refer to our manuscript.
2023-02-05
Western Region
Central Region
Atlantic Region
Outbreak Forecast
Outbreak Forecast
Various data sources could also be used for further surveillance of avian influenza outbreaks. Generally, a rise in Google trends, Google news, and the number of Reddit posts means that the number of avian influenza cases may be increasing in wild or domestic birds. In contrast, a rise in air quality parameters, i.e. CO and IV-index, suggests a reduced likelihood of an avian influeza outbreak.