Improving Citizen Science Games through Open Analytics Data

Game developers make heavy use of "metrics" (or "analytics") in order to understand how their games are being played by the public. Through recording players' actions within a game, designers can identify weaknesses in the game design, such as sections of the game that are too difficult, or never explored. Meanwhile, a genre of “scientific games” has emerged, in which players contribute to the advancement of science through their actions in a game, learn scientific concepts, or both. Although most scientific games incorporate metrics, the data is meant only for the developers and is rarely shared with the public. This represents a missed opportunity for the public to learn from the data that they and their fellow players are generating. Inspired by the belief that anyone can contribute to science, we have created an open source game analytics service called "RedMetrics" in which all data gathered is freely and immediately available online. RedMetrics can gather data from any platform (web, PC, console, etc.) and stores it on an open repository. The data is available via a web API as well as a web application. To ease integration, we provide interfaces for the popular game engine Unity as well as for the web browser. In this paper, we demonstrate RedMetrics through a case study of Hero.Coli, a game to learn synthetic biology. We study the progression of players through the game and use RedMetrics to identify bottlenecks in the game design that hinder learning.