User guided movement analysis in games using semantic trajectories

Understanding how players navigate through virtual worlds can offer useful guidance for map and level design of video games. One way to handle large-scale movement data obtained within games is by modelling movement as a sequence of visited locations instead of focusing on raw trajectory data. In this paper, we introduce a visualization approach for movement analysis based on semantic trajectories derived from a user-guided segmentation of the game environment. Based on this concept, the visualization offers an aggregated view of movement patterns together with the possibility to view individual paths for detailed inspection. We report on a user study with six experts from the game industry and compare the insights they have gleaned from the visualization with feedback from players. Our results indicate that the approach is successful in localizing problematic areas and that semantic trajectories can be a valuable addition to existing approaches for player movement analysis.