This report, just released on PLoS One, is one of the most effective uses of social media, Twitter in this instance, to analyze health characteristics that I have seen. They produce a ‘Lexicocalorimeter’ using sophisticated phrase analyses to produce broad calorie-in and calorie-out categories, and also use the ratio of these two parameters. Impressively, the ratio shows, on a state level, correlation coefficients of over 0.75 with survey-based data on hypertension, diabetes, physical activity, and quality of life. Approaches like this, with rigorous use of data will eventually produce better and faster ways to analyze important health and wellness characteristics for our communities. They will already tell us some things that we already know: the most common food mention was pizza, and the most common activity mention was watching TV or a movie. Just one reason why diabetes prevalence has been growing.
See full article here