This document summarizes Clare Llewellyn's research on analyzing argumentation structures in user-generated online content. It discusses identifying topics and core points in discussions, classifying text spans into argumentation relationships like claims and counterclaims, and using machine learning techniques to automatically classify tweets into an argument structure. Key steps involve topic modeling, rule-based extraction of argument components, and training classifiers on feature sets like n-grams and part-of-speech tags to identify argumentative moves. The goal is to structure unstructured online discussions to better understand persuasive conversations on the web.