PLATO: A visual analytics system for gameplay data

During the last decade the game industry has evolved into a highly competitive market. This has prompted game developers to seek ways to increase the quality of their games which in turn is, to a large extent, dependent on the quality of the player experience. In addition to adapting qualitative evaluation methods, developers have therefore started to use instrumentation techniques to unobtrusively collect large amounts of data of player behavior over time. This creates the need for adequate analysis tools in order to explore and make sense of the data. In this paper we present PLATO, a visual analytics system for time-dependent and multidimensional gameplay data. Gameplay is formally represented as a graph which gives us the advantage of a general representation and makes the tool applicable to a wide variety of games. Moreover, doing so enables us to draw upon a large number of graph algorithms. PLATO integrates techniques for subgraph matching, pathfinding, data comparison, and clustering as well as several visualization techniques. We demonstrate the utility of our system by analyzing real world data from a massively multiplayer online role-playing game, a team-based first person shooter, a puzzle game, and a platformer. (C) 2013 Elsevier Ltd. All rights reserved.