Johns Hopkins Uses Semantic Analysis of Tweets to Identify Health Trends

Johns Hopkins is using semantic analysis to sift through tweets for information about health trends. According to one article, the effort “gives some clues as to how health marketers, especially, can take advantage of consumer sharing habits to inform key decisions – particularly around media spend, shelf placement, and other choices whose optimal resolution can vary significantly between – and even within – local markets. Researchers at that JHU’s School of Public Health analyzed 2 billion public tweets, from May 2009 to October 2010, to determine whether semantic analysis could unearth meaningful information about public health trends.”

It continues, “The results were remarkable: plotting the overall Twitter chatter around the word ‘flu’ and the actual reported flu rate for a given week according to the Centers for Disease Control indicated that the two metrics are highly correlated (see image above). So, for influenza, at least, the self-reported flu rate as indicated by social sharing closely mirrors the flu rate as measured by more typical epidemiological methods. I could imagine that, for other familiar ailments (think colds, stomach viruses, and headaches) the chart might look the same.”

Read more here.

Image: Courtesy Johns Hopkins