Abstract:
|
Twitter provides an open, easy to obtain data source which provides information about the major topics of discussion within communities. For Biosurveillance, it provides information about how the public is reacting to an event, what rumors are being propagated, what the reaction is to politicians, celebrities, and the media reporting on the event. To some degree it can be used for early detection of outbreaks. However, Twitter data has several aspects that need to be considered in any analysis. First, one does not have full access to all the data for free; the API provides a subset of the data. Second, not everyone uses Twitter, and the demographics of the users can be different in different regions of the world. Demographics and usage changes in time, and these changes may not be well understood. I will discuss the issues of early detection and situation assessment as the event progresses, illustrating this with several different disease outbreaks that have been in the news recently. I will also discuss methods for assessing and classifying different "types" of Twitter accounts such as robots, news and medical information sources, and various types of general public accounts. This
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.