Since the rise of social networks, daily exposure to fellow users’ opinions and comments has become a constant in many people’s lives. Research has shown the impact of even the most passive type of online user engagement, namely reading comments under a news item, can affect the perceived quality of the article, as well as change users’ perception of the public opinion of a certain issue. In order to automatically detect the reasoning underlying users’ opinions on political issues of public interest, this project aims to research a new methodology for argumentation mining of social media data in Dutch and English. We propose a novel framework and methodology to automatically extract topics, stance and argument structures from (political) social media data. A varied social media corpus will be collected and annotated. The insights gained from our pilot annotation study for Dutch will form the basis of a machine learning approach to automatically extract topics, stance and argumentative components and relations from our corpus. The results of our research will push the state-of-the-art in natural language understanding and help communication scientists and social scientists to formulate new hypotheses regarding the evolution of digital politics.