This project explores player behavior in The Resistance by focusing on non-verbal clues, specifically eye movements and visual attention throughout the game. Everything used has been extracted from "https://snap.stanford.edu/data/comm-f2f-Resistance.html"
We applied social computation techniques to extract meaningful metrics such as betweenness centrality, graph diameter, modularity, and clustering coefficient with a special focus on similarity, which plays a key role in generating embeddings to facilitate deeper analysis.
To begin, player embeddings were created in the notebook generation.ipynb, based on similarity of behavior. The visualization file displays a community-based representation of these embeddings. In the analysis file, various metrics are calculated to highlight differences between groups, helping to identify where the most likely deceivers may be hiding.
- Daniel Moraleda Sánchez
- Daniel Navarro Puche