Global Sentiment is a research programme using social-media text at planetary scale to measure how the environment — air pollution, weather, climate and the urban fabric — shapes expressed human sentiment. It brings together the MIT Sustainable Urbanization Lab and the Center for Geographic Analysis at Harvard University, with methods rooted in geospatial big data and computational social science.
The core premise: people's moment-to-moment expressions, aggregated over millions of geolocated observations, form a measurable signal of how environments feel — a signal we can relate to pollution, temperature, and the built environment.
Across studies we have shown that air pollution lowers expressed happiness, that daily weather moves sentiment for tens of millions of people, that rising temperatures degrade global sentiment unequally, that the pandemic altered how the world expressed emotion, and that even the wavelength of urban night-time light modulates how people feel. Five of the programme's landmark papers are listed below.
Collaboration
The project is developed in partnership with the MIT Sustainable Urbanization Lab and the Center for Geographic Analysis, Harvard University. The public portal for the programme lives at globalsentiment.mit.edu.
