The study, published in online journal PLoS One, looked at developing a model of “the contagious spread of information in a global-scale, publicly articulated social network”.

Results were based on the analysis of 6 months of Twitter data recorded in 2009 and based on a network containing 40 million Twitter users worldwide (connected by 1.5 billion directed relationships).

The authors identified “sensor” groups – individuals at the centre of a network – and looked to test the hypothesis that collecting information from these individuals could be used to detect contagious outbreaks before they happen in the population at large. It was found that a group of 50,000 was the optimal size of sensor group for predicting virality.

The results suggested that these sensor individuals could predict the rise of certain hashtags up to 20 days before their less well-connected counterparts. The authors proposed that this could inform the use of social media to help predict “important phenomena” such as flu outbreaks, global mood patterns, or movements in the stock market.