In this project, Jörg Friedrichs and a small team of researchers apply machine learning and sentiment analysis to social media and other data to investigate the political usages of fear and anger. An article about fear-anger contests in the context of the 2016 Brexit referendum and the 2016 presidential election of Donald Trump has appeared, showing that established and populist actors outbid one another in expressing fear and anger. Populists express more anger, whereas establishment or governmental actors express more fear. Thus, Brexit was a contest between “project fear” (Remain) and “project anger” (Leave). In the 2016 presidential campaign, the Democrats led on both fear and anger but this did not win them the election. Findings are based on machine learning and time series analysis of Twitter data. Following up on this first publication, Jörg and his team are working on another article about the political usages of fear and anger in the context of COVID-19.