Maps of Heart Disease Rates

Penn Team uses Machine Learning and Twitter to Predict Heart Disease Risk

A Penn team, including IBI faculty member Lyle Ungar, won Scientist of the Year at the 2015 Philly Geek Awards. The team demonstrated that Twitter can serve as a dashboard indicator of community psychological health and can predict rates of coronary heart disease.


Using natural language processing, sophisticated artificial intelligence techniques that allow computers to interpret human language patterns, the researchers demonstrated that Twitter can capture more information about heart disease mortality risk than the traditional risk factors. The research provides additional evidence of the relationship between mood and physical health. They found that expression of negative emotions in Tweets was associated with higher heart disease mortality risk. Positive emotions like optimism and excitement were associated with lower risk. More information on the study can be found here.  The study was conducted as part of The Penn World Well-Being Project.