Towards real-time influenza surveillance: exploiting twitter data

Authors: Tian Wen*, SUNY - Buffalo, Ling Bian, SUNY - Buffalo
Topics: Applied Geography
Keywords: twitter, machine learning, health
Session Type: Poster
Day: 4/5/2019
Start / End Time: 8:00 AM / 9:40 AM
Room: Lincoln 2, Marriott, Exhibition Level
Presentation File: No File Uploaded


Real time Influenza surveillance is of paramount importance to public health and for disease prevention and control. Although many data sources have been exploited to conduct influenza surveillance, most of them cannot support real-time surveillance or are not easily accessed by the public. Twitter data are valuable for the intended surveillance as it provides real-time data in large quantities on daily basis, and the data are open to the public. This research explores the possibility of real-time influenza surveillance in an urban area during the 2016-2017 flu season using Twitter data. The flu-relevant and flu-irrelevant tweets were selected from sample tweets according to the literature. A support vector machine (SVM) approach was applied to identify flu-relevant tweets during the flu season as an indication of flu prevalence in the area. Results show that the twitter-based flu prediction coincides with the published prevalence of flu in terms of the general trend.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login