Do they play as intended? - Comparing aggregated and temporal behavioral analysis in a persuasive gamified system

Abstract

Gamified systems nurture an ulterior goal set by their designers (e.g., a positive behavioral change). Behavioral profiling allows understanding whether users play as intended and reach such a goal. Analyzing in-game behaviors can also highlight unexpected interaction patterns or unengaged users. Current logging systems can track and store any in-game action. However, such high-dimensional data should be carefully processed to retain relevant knowledge while filtering unnecessary noise. Analysts can either aggregate data into a single data point per player or maintain temporal information. This study compares aggregated and temporal behavioral analysis conducted on a gamified system, promoting sustainable mobility (Play&Go). Results show how, in Play&Go, aggregated analysis conveys information on long-term winning strategies, whereas temporal analysis describes short-term strategies. Additionally, studying the temporal evolution of players' behaviors emphasizes a sharp division among engaged and unengaged users. We show how aggregated and temporal analysis hold a complementary view of players' experiences.