Fits Like a Game: A Multi-criteria Adaptive Gamification for Collaborative Location-Based Collecting Systems

Collaborative location-based collecting systems often struggle with engagement because static, uniform gamification strategies fail to account for diverse user motivations and project requirements. To address this, a new adaptive framework utilizes a multi-criteria recommendation system to tailor game challenges to individual participants. By balancing personal preferences with broader project objectives—such as the need for comprehensive spatial and temporal data coverage—this approach moves away from rigid, one-size-fits-all models toward a dynamic, personalized experience that enhances both user participation and overall data collection efficiency.

The technical architecture relies on the analysis of behavioral atoms, which are processed through K-means clustering to generate user traveling behavior time series. This data facilitates collaborative filtering, allowing the system to predict and assign game element scores based on the preferences of similar users. The framework further refines these recommendations by incorporating a feedback loop that adjusts user profiles based on their specific selections. Preliminary evaluations using adapted datasets have yielded a precision rate of 0.9, demonstrating the high accuracy of this predictive model in identifying effective gamification elements.

While current validation remains confined to offline metrics, the framework establishes a robust foundation for future real-world applications in fields like citizen science. By employing a non-intrusive disaggregation function to model user behavior, the system minimizes friction while maximizing engagement. Ongoing development focuses on transitioning from offline testing to live environments, where the integration of dynamic difficulty scaling and real-time reward adjustments will further optimize the synergy between individual user activity and large-scale collaborative data collection goals.