A person’s tweets are a window into their personality. So much so, that researchers from the World Well-Being Project at the University of Pennsylvania were able to map personality traits onto thousands of counties in the United States using a collections of tweets.
The researchers first built a model based on the myPersonality database, in part based on the results of personality tests taken through Facebook, including the so-called “big five” (which measures neuroticism, agreeableness, extraversion, openness to experiences, and conscientiousness), and in part based on the Facebook posts of those who took the personality tests. Kokil Jaidka, a postdoctoral research fellow at the project, explains that these data were then used to create a model that analyzes an individual’s personality according to the language they use on social media, including Twitter.
The well-being map of the US shows the personality not of individuals, but counties, based on 1.53 billion tweets. Each county included has at least 30,000 tweets (a minimum of 100 people with at least 30 tweets each.) The most neurotic places are charted based on their z-score (how many standard deviations they are from the mean neuroticism.)
There are potential health implications from these findings—research from the team has found that the same language that predicts neuroticism also predicts heart disease mortality.