We leveraged exogenous variation in weather patterns across geographies to identify social contagion in exercise behaviors across a global social network. We estimated these contagion effects by combining daily global weather data, which creates exogenous variation in running among friends, with data on the network ties and daily exercise patterns of ~1.1M individuals who ran over 350M kilometers in a global social network over five years. Our analysis shows exercise is socially contagious and that its contagiousness varies with the relative activity of and gender relationships between friends. Less active runners influence more active runners, while the reverse is not true. Both men and women influence men, while only women influence other women. While the Embeddedness and StructuralDiversity theories of social contagion explain the influence effects we observe, the Complex Contagion theory does not. These results suggest interventions that account for social contagion will spread behavior change more effectively.