Play-Graph: A methodology and visualization approach for the analysis of gameplay data

Instrumentation to automatically log information - so called gameplay metrics - about the player-game interaction has become an important tool for analyzing player behavior in games. However, due to the usually large amount of gathered data, analysis of the collected data can be challenging. Visualizations have become a promising addition to statistical techniques to explore and better understand the data. In this paper we build upon our previous work on gameplay analysis and introduce the concept of a play-graph as a way to formally describe and visualize gameplay data. Difference graphs are used to depict the changes between two different datasets - a relevant but currently neglected task in gameplay analytics. Furthermore, we address issues in regard to the visual representation of our former prototype. Data obtained from a puzzle game, a team-based shooter, and a MMORPG are used to illustrate the concepts.