Network based model of social media big data predicts contagious disease diffusion

Area of Science:

• Epidemiology
• Computational Social Science
• Network Science

Background:

• Predicting disease outbreaks is crucial for public health.
• Social media big data offers novel opportunities for real-time disease surveillance.
• Understanding disease diffusion across geographies is a significant challenge.

Purpose of the Study:

• To analyze social media micro-blogs and geographical locations to understand disease spread.
• To enhance forecasting of future disease outbreaks using big data.
• To develop a generalized solution for tracking contagious disease diffusion across geographies.

Main Methods:

• Utilized Twitter data as the social media data source.
• Employed a novel network-based model for disease spread prediction.
• Used influenza-like illnesses (ILI) as the epidemic and US states as geographical locations.

Main Results:

• Flu-related tweets showed strong alignment with Centers for Disease Control and Prevention (CDC) ILI data (p < 0.049).
• The network model provided more accurate ILI activity estimates compared to airline traffic models.
• Accurate predictions of upcoming ILI activity (p < 0.04) and 76% accuracy in predicting flu diffusion across states were achieved.

Conclusions:

• The developed model offers a generalized approach for analyzing disease diffusion across geographies using social media data.
• The methodology can be adapted for various social media platforms, diseases, and geographical scales.
• This research advances the field of real-time disease surveillance and outbreak prediction.