In this project, I will focus on various aspects of online communication, particularly on under-researched modularities such as the audio and visual information in videos. Initially, I will concentrate on detecting emotional changes and persuasiveness in audio recordings, such as those found in podcasts. The current goal is to analyse changes in voice during the radicalisation process, such as in the alt-right pipeline. In addition, the project aims to understand the impact of different modularities on phenomena such as polarisation, radicalisation and disagreement, but also uses modularities as a tool for coordination detection.