Description:
We created a unique data set based on social media data by collecting and geo-localising all the tweets of 54 thousand Swedish citizens from January 2019 to June 2019. This allows us to construct an attractive individual-level measure of preferences for pro-environmental behavior. We demonstrate this by using our measure in two applications. We first document a subjective well-being gap between individuals with and without green preferences, using the average sentiment scores in tweets as a proxy of individuals' subjective well-being. We then investigate the existence of a gender gap in green preferences and the propensity to act for the environment, relating our measure to publicly available data on electric and hybrid car registrations and political support for environmental policies in Sweden.