A team of linguists, information scientists and political scientists embarks on the question when political negotiations are successful and why they are successful. In particular, we want to develop automatic tools that analyse political discourse, allowing to draw conclusions on its effectiveness. The motivation behind this is that public mega-projects repeatedly create conflicts between governments and the public sphere and their realization has become a incalculable risk for political decision makers. Therefore we want to investigate the factors that make political communication successful.This complex task can only be tackled using an innovative combination of methods from different areas of research. These methods include 1) a deep and detailed linguistic processing of real mediation processes to generate an abstract representation of communication; 2) a shallow, statistical analysis of text to detect common patterns in negotiations and 3) the development and employment of visualization tools which identify patterns of communication at-a-glance.