Modeling Individual and Team Behavior through Spatio-temporal Analysis
Sabbir Ahmad, Northeastern University,
Andy Bryant, Northeastern University,
Erica Kleinman, Northeastern University,
Zhaoqing Teng, Northeastern University,
Truong-Huy D Nguyen, Fordham University,
Magy Seif El-Nasr, Northeastern University
This paper introduces a new methodological platform for identifying, understanding, and modeling player behaviors called Interactive Behavior Analytics (IBA). Targeted towards researchers, game designers, and developers, IBA expands our understanding of individual and team behaviors, strategies, and tactics facilitating innovation in game analytics and game design, particularly expanding the techniques used for game personalization, game user research, and the development of game guides and NPCs (Non-Player Characters).
Unlike current methods that rely purely on machine learning, IBA takes on a qualitative methodological approach using quantitative game data collected from play sessions with embedded and integrated visualization systems: StratMapper and Glyph, a labeling mechanism, and abstraction algorithms. The methodology is packaged in a seamless interface to facilitate a human-in-the-loop approach to knowledge discovery from game data. The human-in the loop feature is key as it allows capturing important contextual and situational variables that are otherwise left unaccounted for with a purely statistical or quantitative technique.
This paper details the methodological approach and demonstrates its utility through two case studies in which it was used to analyze two games: BoomTown, a game developed by Gallup, and DotA 2. Through the case studies, it is apparent that IBA pushes the current state of the art, giving us a tool to gain a deeper look into players’ problem-solving behaviors, decisions, and strategies through visualization and analysis by integrating contextual details and sequences of behaviors, rather than aggregate variables and integrating time. For more details on IBA and the example case studies analyses and results, we invite you to read our paper.
Contact author: Sabbir Ahmad
CHI PLAY session:
Analyzing & Visualizing Player Behavior
Friday, 25 October 2019, 14:00–15:30
Please feel free to also write comments or questions for the authors in the space below!