Recently, many news articles had points, as to how a social networking site affects a user and the world negatively, but Marcel Salathe, assistant professor of biology, Penn State University, has noted a positive instance with Twitter. He conducted a unique analysis of how the popular mircroblogging site records the spread of an epidemic of its users’ receptivity to new vaccines.
Later on, Salathe tracked the process of how the users’ thinking correlated with vaccination rates. Additionally, he even noted the processing of how microbloggers with the same negative or positive feelings seemed to influence others.
This study performed by Salathe can be assumed as the first ever study relating to social-media sites’ affecting and reflecting disease networks. The journal of Public Library of Science Computational Biology even reports that this method can be ideally utilized even to track other diseases.
Salathe noted that there were two reasons for him choosing Twitter. First reason was that unlike the contents of Facebook, Twitter messages or “tweets” are considered public data. Moreover, according to a Penn statement, one can easily “follow” or track the tweets of anyone else. Secondly, the microblogging site is a perfect database for one to learn about people’s sentiments. Salathe explained, “Tweets are very short — a maximum of 140 characters.” he added, “So users have to express their opinions and beliefs about a particular subject very concisely.”
Salathe started moving towards his goal, with a number of tweets, which counted up to 477,768. These tweets had keywords and phrases, which were related to vaccination. Next, he worked on tracking sentiments of a user about a particular new vaccine for combating H1N1, a virus strain responsible for swine flu.
The collection took a long time as it began in August 2009 and it continued up to January 2010, when news of the new vaccine first was made public.
The collection seemed to be an easier task, but sorting was a hassle, as Salathe noted that he partitioned a random subset of about 10 percent and asked Penn State students to rate them as positive, negative, neutral or irrelevant. For the lingering thoughts, he had an example which would ease the confusion. It was, a tweet expressing a desire to get the H1N1 vaccine would be considered positive, while a tweet expressing the belief that the vaccine causes harm would be considered negative.
This meant that if there was a tweet, which concerns a different vaccine, for example, the Hepatitis B vaccine, it would be irrelevant for the cause. The tweets were finally sorted out and the final tally was 318,379 tweets, which was after the elimination of the irrelevant ones. These tweets were expressing either positive, negative or neutral sentiments about the H1N1 vaccine.
Additionally, as there is an inclusion of location in the profiles of Twitter users, Salathe was helped further to classify the sentiments expressed by the US region.
Salathe concluded noting that he was able to find definite patterns as he said, “The assumption is that people tend to communicate online almost exclusively with people who think the same way. This phenomenon creates ‘echo chambers’ in which dissenting opinions are not heard.”