“just setting up my twttr,” social media entrepreneur Jack Dorsey typed into a website. Ten years later, Twitter is the place where much of the world “talks to itself” in 140 characters or less.
A July 15, 2006, news story is credited with bringing Twitter into the public’s consciousness, and since then it has become a fertile ground for researchers interested in tracking social, cultural, and political trends, including topics such as disease outbreaks, the dynamics of campaigns, and consumer preferences.
Dorsey and his cofounders chose the name “twitter” because it described “a short inconsequential burst of information.”
And yet, Twitter is anything but “inconsequential” in terms of data science research and its applications.
Using machine learning, natural language processing and other data science techniques, researchers leverage Twitter’s 300 million followers and the hundreds of millions of tweets they post each day as a kind of distributed sensor network, where each person observes and reports on some aspect of the world, says Henry Kautz, the Robin and Tim Wentworth Director of the Goergen Institute for Data Science at the University of Rochester.
“Each report is very noisy, but the aggregate results can be reliable,” Kautz says. “The approach can be used for health, environmental protection, public safety and many other applications.”
For example, at the University of Rochester:
- Kautz and his team have used Twitter to reliably identify and track people with symptoms of flu and food poisoning , offering a much quicker way for health officials to respond to disease outbreaks and even to forecast when and if a specific individual will fall ill. His team has also used Twitter to analyze patterns of alcohol consumption.
- Jiebo Luo, associate professor of computer science, and his team have tracked the Twitter followers of Donald Trump, Hillary Clinton, Bernie Sanders and other presidential candidates since September to better understand the dynamics of the campaign. In one paper, for example, they found that Trump followers responded more favorably when he “bashed” Democrats rather than Republicans. And after Trump stated that “the only thing (Hillary Clinton’s) got going is the woman’s card,” the researchers found that women were more likely to follow Clinton and less likely to “unfollow” her.
- Luo and his students have also used Twitter to correlate the various language styles and interests people express in tweets to their different occupations, and to predict people’s jobs accordingly.
- Huaxia Rui, an assistant professor at the Simon School of Business, and his colleagues, have shown how business managers can glean important clues about the popularity of their products and even forecast future sales through careful analysis of Twitter traffic. They also discovered that airlines are most likely to respond to customers who complain via Twitter if they have a larger following.
- Las Vegas Health Department recently field tested the nEmesis app developed by Kautz and his team to connect food-poisoning-related tweets to restaurants. The researchers found that the tweet-based system led to citations for health violations in 15 percent of inspections, compared to 9 percent using the traditional random system, resulting in 9,000 fewer food poisoning incidents and 557 fewer hospitalizations during the course of the study.
Read more about data science at the University of Rochester.
