G-Player: Exploratory Visual Analytics for Accessible Knowledge Discovery

G-Player is a spatio-temporal visual analytics prototype designed to make game telemetry analysis accessible to non-experts. The work addresses a gap in existing practice: aggregated statistical tools lose the individual behavioral detail that designers need, while individual trace viewers fail to scale to the hundreds of thousands of logged data points generated per user. By combining heat maps, movement plots, and Boolean operators, G-Player allows analysts to explore high-dimensional, time-varying player data in both space and time.

The system's central contributions are space-time constraints and Boolean-compound heat maps, which allow users to filter and compare behavioral patterns with precision. The authors demonstrate these features using a pickpocketing example drawn from the VPAL case study, a Fallout: New Vegas modification that produces extensive per-user logs. This grounding in a concrete, large-scale dataset gives the tool's design choices practical context.

Evaluation consisted of one-on-one sessions with specialists from Ubisoft and Unity, who completed three analytical tasks and compared G-Player against their own tools, Ubisoft's DNA tool and Unity's Heatmap plugin. The authors acknowledge selection, social-desirability, and self-defense biases in this setup and argue that these partially offset one another. Results from a feature comparison indicate that G-Player matched or exceeded competitors in data handling and dimensionality, though it trailed in usability, a limitation the authors identify as an area needing further work.

Overall, the expert reception was favorable, with both Ubisoft and Unity practitioners praising the tool and signaling intent to incorporate its playback, spatiotemporal constraint, and Boolean operation features into their own products. This suggests the approach has practical relevance for industry game analytics beyond the academic prototype.