Uncovering Learning Patterns in Collaborative Museum Games
This site documents a multi-phase research study investigating individual learning patterns within collaborative, multiplayer game environments in museum settings. Using computational behavioral analysis (t-SNE spatial embedding, critical change points detection), statistical validation methods (mixed ANOVA, trajectory progression analysis), and integration of the statistical analysis with machine learning techniques, this work reveals the relationship between strategic convergence and learning progression during collaborative gameplay.
Research Focus
Understanding and supporting collaborative learning through multi-method analytics in authentic educational environments.
Core Questions
- When and how do critical behavioral shifts occur during individual learning within collaborative gameplay sessions?
- What computational methods can reliably detect and characterize these learning transitions?
- What distinct individual behavioral patterns emerge during collaborative learning sessions?
- How can integrated spatial-temporal analytics inform evidence-based design for collaborative learning environments?
Research Scope
- Data: 1.5M+ game logs on learning analytics (1500+ sessions, 3000+ players), prelim focused 90 sessions, multi-method validation pipeline
- Methods: Learner change points detection, spatial pathways, statistical progression analysis
- Implications: Evidence-based design principles for informal learning environments
- Framework: Systematic approach from behavioral understanding to learning support design
Explore the Research Journey
- 1. Exploratory Analysis
Initial deep dives into change points, player modeling, and spatial trajectories.
Emerging Insights
This work-in-progress explores a range of methods and analytical strategies that may be useful for researchers studying informal learning, spatial interaction, and game-based environments. While still under development, these approaches show potential for informing future research and design.
Analytical Approaches Explored
- Change Point Detection in Play
Identifies shifts in learner engagement or strategy using z-score analysis across behavior windows. - Spatial Behavior Modeling
Reconstructs learner pathways to explore navigation styles and exploration patterns. - Predictive Modeling of Session Outcomes
Uses logistic and multinomial models to examine how early behavioral indicators relate to success metrics. - Trajectory-Informed Hypothesis Generation
Links behavior patterns to evolving research questions, demonstrating a possible model for exploratory-to-confirmatory transitions.
Methodological Contributions
- Complete Analysis Pipeline
Demonstrated end-to-end process from raw logs to actionable insights - Replicable Methodology
Framework adaptable to other collaborative learning contexts - Visualization Suite
Multiple chart types for spatial, temporal, and integrated analysis