Psychological Language on Twitter Predicts County-Level Heart Disease Mortality

Abstract

Hostility and chronic stress are known risk factors for heart disease, but they are costly to assess on a large scale. We used language expressed on Twitter to characterize community-level psychological correlates of age-adjusted mortality from atherosclerotic heart disease (AHD). Language patterns reflecting negative social relationships, disengagement, and negative emotions—especially anger—emerged as risk factors; positive emotions and psychological engagement emerged as protective factors. Most correlations remained significant after controlling for income and education. A cross-sectional regression model based only on Twitter language predicted AHD mortality significantly better than did a model that combined 10 common demographic, socioeconomic, and health risk factors, including smoking, diabetes, hypertension, and obesity. Capturing community psychological characteristics through social media is feasible, and these characteristics are strong markers of cardiovascular mortality at the community level.

Exploring Small, Confirming Big: An alternative system to The New Statistics for advancing cumulative and replicable psychological research

“…Sentiment analysis of social media data is becoming increasingly common (e.g., Schwartz et al, 2013), and could range from utilizing relatively simple pre-programmed psychological language dictionaries (e.g., Pennebaker, Chung, Ireland, Gonzales, & Booth, 2007), to complex machine learning/data-driven language categories (Eichstaedt et al, 2015).…”

Language, demographics, emotions, and the structure of online social networks

“…Recognizing this limitation, our conclusions only apply at the group level and not at the level of individual behavior of LA residents. In the same vein as how Twitter data can be used to identify group effects on heart disease mortality [36], our analysis identifies relations between properties of groups of people. While recent research opens the possibility to how to reweight Twitter metrics across demographic sections [37], evaluating the external validity of social media metrics requires an interdisciplinary effort beyond the scope of this contribution from data science.…”

Tweeting About Mental Health: Big Data Text Analysis of Twitter for Public Policy

“…A lexical approach is first applied to get a basic polarity of a text (e.g., tweets).Then, any opinion words (those with a non-zero emotional content according to the lexicon) are stripped out, and a classifier is applied to predict positive and negative tags based on the remaining domain specific features. In the case of Twitter, domain specific features would be emoticons, abbreviations, and 119 (Pawar, Shrishrimal and Deshmukh) 120 (Kim et al, 2009) 121 (Pawar, Shrishrimal and Deshmukh) 122 123 (Eichstaedt et al, 2015) 124 (Liu and Zhang, 2012) 125 (Ji et al, 2015) 126 (Mohammad, Kiritchenko and Zhu, 2013) misspellings 127 . This approach allows sophisticated linguistic analysis to be combined with a problem specific classifier, creating a richer model.…”