Abstract
The rich data generated and read by millions of users on social media tells what is happening in the real world in a rapid and accurate fashion. In recent years many researchers have explored real-time streaming data from Twitter for a broad range of applications, including predicting stock markets and public health trend. In this paper we design, implement, and evaluate a prototype system to collect and analyze influenza statuses over different geographical locations with real-time tweet streams. We investigate the correlation between the Twitter flu counts and the official statistics from the Center for Disease Control and Prevention (CDC) and discover that real-time tweet streams capture the dynamics of influenza cases at both national and regional level and could potentially serve as an early warning system of influenza epidemics. Furthermore, we propose a dynamic mathematical model which can forecast Twitter flu counts with high accuracy.
Original language | English (US) |
---|---|
Article number | 189 |
Journal | Journal of Medical Systems |
Volume | 40 |
Issue number | 8 |
DOIs | |
State | Published - Aug 1 2016 |
Keywords
- Geo-tagged twitter stream
- Influenza
- Partial differential equation modeling
- Regional level
ASJC Scopus subject areas
- Medicine (miscellaneous)
- Information Systems
- Health Informatics
- Health Information Management