For example, in a study of 148 million Twitter messages (tweets) across 1,347 counties in the United States, language patterns reflecting negative social relationships, disengagement, and negative emotions—especially anger—emerged as risk factors for premature death;

