MEXCA

The aim of MEXCA is to capture the emotional appeals of politicians. In election debates, tv interviews and parliamentary sessions politicians make appeals to our emotions: they want us to be angry about an issue, proud of our country, or fearful about the future. In making these appeals they use their face to make expressions, they choose particular words and also use the tone of their voice. Not all of this is deliberate and controllable. Yet what the face, the voice, and the text communicate impacts the listeners. The face, the voice and text are different modalities that communicate emotions. MEXCA is software that takes a video of an election debate, a tv interview or discussion, or a parliamentary session and it identifies the multimodal emotion expression of the people in the video. It automatically assigns video and audio segments to the same person, and uses external tools to analyze changes in facial expressions, tone in voice and sentiment in text.

MEXCA is currently open-source software downloadable from Github (https://github.com/mexca/mexca). It is currently deployed in two projects: (1) a large-scale analysis of 78 German election debates in the period 1997-2021; and (2) a series of focus-groups conversations around climate change. At this stage the data analysis is still ongoing.

The software is open for other people as well. Following three workshops the intended audience primarily consists of social scientists.