A biology professor is using Twitter to track how receptive the public is to vaccines like the H1N1, observing how many people perceived the vaccine positively, negatively, and neutrally around the United States.
Marcel Salathe examined 477,768 tweets from August 2009 to January 2010 which contained vaccine-related keywords. He then narrowed this down to only those tweets pertaining to the H1N1 vaccine in particular.
Salathe chose Twitter – as opposed to Facebook or another social network – because of two reasons: it is public, and it expresses sentiment.
Sentiment analysis is something of a buzz word in social media circles these days, but Salathe didn’t rely on over-hyped software to test how people felt about the H1N1 vaccine. Instead, he pared the tweets down to just 10 percent and handed them out to Penn State students to comb through.
Students were tasked with rating the tweets as positive (“Looking forward to getting my vaccine this afternoon”), negative (“I hear the H1N1 vaccine causes more problems than H1N1 itself”) and neutral (“H1N1 vaccinations at school today”).
From this data, Salathe used a computer algorithm to measure sentiment across the US, using people’s Twitter bio location to pinpoint where they were tweeting from.
He found some pretty interesting correlations between what people tweet about the vaccine and what they actually do: New England had the highest percentage of positive sentiment towards the vaccine, and it also had the highest vaccination rate in the country.
There is hope that this method can be used to track future vaccines and even diseases or epidemics around the US.