Medical researchers use data including income, education level and other demographics to predict the occurrence of heart disease in populations. Now, researchers at Penn's World Well-Being Project (WWBP), which is part of its Positive Psychology Center, have found a new tool.
"The language patterns on Twitter contain markers for all these demographic variables," explains Johannes Eichstaedt, a founding research scientist of the WWBP.
"When you're trying to predict rates of heart disease among communities, the standard approach is to use things like income, education, and levels of obesity and hypertension. You give a good statistical algorithm all that information, and it does a pretty good job of predicting heart disease," he says. "It's not that Twitter captures a completely different set of information; it seems like Twitter absorbs all this other information."
Eichstaedt will speak about the project at the Penn Science Café, at World Café Live. The original date for the talk, Tuesday, October 7th at 6 p.m. has been cancelled; the rescheduled date is to be determined.
Support provided by