Public deliberation forums produce copious amounts of unorganized textual data reflecting diverse viewpoints. Visualization can serve as a valuable tool for understanding the relationships between policy preferences. We introduce a unique approach to visualizing opinion data gathered from the Polis online platform in which LLMs are used to generate positions and structure the data into argument maps. Each AI-generated position is supported by human-generated statements, providing a more meaningful organization of Polis’s opinion data. We believe that these argument maps can provide policymakers with easy-to-understand visualizations that summarize public sentiment.