Social media chatter reveals county-level male v. female exercise patterns

Combining machine learning with statistical modeling, researchers have analyzed tweets to figure out who’s doing more to keep fit—men or women—along with exactly how the sexes are exercising and approximately where they live.

The study behind the intriguing, if potentially unsettling, science is running in BMJ Open Sport & Exercise Medicine.

Lead author Nina Cesare, PhD, of Boston University and colleagues say the innovation might help public health officials “identify changes in health behaviors at small geographical scales” and “design interventions best suited for specific populations.”

The team gathered Twitter data from 1.4 million tweets posted by 481,000 people (55.7% men and 44.3% women) over a one-year period in more than 2,900 U.S. counties.

Applied to this massive dataset, their machine learning/statistical modeling formula showed walking to be the overall most popular physical activity for both men (15.9%) and women (18.7%) across the country.

Gym-based activities had lower yet similarly close rates for the two sexes, meaningfully mentioned by 4.68% of men and 4.13% of women.

Within the gym, though, men went for CrossFit more than any other option (14.91%) while 26.66% of women drove yoga to the top.

As for regional differences, the West had the most women exercising with intensity, the Midwest had the most intensive exercise among men, and the South had the biggest gap in exercise intensity between men and women.

The researchers further found that indicators of sedentary lifestyle tended to be lower in areas with high levels of activity, possibly confirming the power of social suggestion, such as adult peer pressure.

Commenting on this finding in a news item posted by the Boston University School of Public Health, senior study author Elaine Nsoesie, PhD, points out that lower-income communities tend to lack health and wellness resources.  

“By understanding differences in how people are exercising across different communities, we can design interventions that target the specific needs of those communities,” Nsoesie says.

Cesare et al. concluded that Twitter is “a useful tool for measuring small-area trends in physical activity, an important risk factor for non-communicable diseases,” with the caveat that its usefulness might vary by sex and region.  

Click here to read the full study and here to read the university’s news item.