Making visible patterns within conversations enables us to see dynamics between participants, the impact of facilitators, and perhaps even quality of the conversation itself. In this project, we leverage natural language processing, machine learning, data visualization, and human-computer interaction techniques to unveil and create metrics for dynamics within conversations in an effort to evaluate quality of conversation. The Jellies are small network maps of each conversation, nodes being participants and edges between them. We can sort the conversations by various metrics, such turn taking distribution or average levels of responsivity. We hope that such visualizations will support conversation designers as they learn, reflect, and drive towards constructive communication.
