Incorporating Learning Analytics in an Educational Game to Provide Players with Information about how to Improve their Performance* J. X Seaton, Sabine Graf, Maiga Chang, & Arta Farahmand School of Computing & Information Systems Athabasca University Edmonton, Canada j.xseaton@gmail.com, sabineg@athabascau.ca, maigac@athabascau.ca, & arta.farhmand@gmail.com Abstract—Educational games aim to balance learning and playing. However, for people to benefit from an educational game, they must be encouraged to play the game often. Providing players with information about how to improve their performance could help in achieving this goal. This paper examines how a learning analytics dashboard can be incorporated into an educational game to encourage players to play more often and continuously. The proposed dashboard provides players with a variety of information such as how their performance and skills change over time. Such information allows players to see their performance and play habits, and find strategies on how to improve their performance, and therefore their learning, in the game. * Keywords-learning analytics; educational game; dashboard; metacognitive skills I. INTRODUCTION A challenge to educational games is that it can be hard for a player to connect the feedback in a game to how they can further improve in the game. As players progress to more challenging problems, often they learn by failing and re- attempting the challenge. Feedback about a failed attempt can obscure how that failed attempt helped the player become better at playing the game. Incorporating learning analytics through visualizing gameplay data, including the player’s scores, position, or decisions made in the game, can provide meaningful information about how the player progressed through the game [1]. For example, the game CMX is an educational Massively Multiplayer Online Role Playing Game (MMORPG) that teaches computer programming [2]. The game incorporates learning analytics by creating reports for instructors about how players are progressing within the game (e.g., how many learning activities students have completed, how many errors they made, etc). Educational games can also use learning analytics to provide feedback to learners. For example, the educational game eAdventure [3] provides students with reports that assess their learning. eAdventure provides learners with information about how they are doing in the game, including how much time they are spending playing the game; the time it took them to finish the game (or a subsection of the game), and their score in the game. * The authors wish to acknowledge the support of Canadian Internet Registration Authority (CIRA)’s Community Investment Program, Natural Sciences and Engineering Research Council of Canada (NSERC), and Athabasca University. While most related works use learning analytics in games with the purpose of assessing and reporting on students’ performance and progress, the aim of this paper is to introduce a learning analytics dashboard that focuses on motivating players and helping them to understand how they can perform better in the game. The dashboard has been evaluated in a proof of concept evaluation with three months of simulated gameplay data. II. LEARNING ANALYTICS DASHBOARD The learning analytics dashboard has been incorporated into an educational game that aims to improve players’ metacognitive skills. The game consists of ten subgames, which each targets improving a metacognitive skill. In the game, players play matches of three subgames against other players. Both players are scored by how they performed individually and against their opponent based on their performance score in each subgame. The player that performs best over a match receives points and the loser loses points, which allow players to be ranked against other players. In addition, a metacognitive skill score is calculated based on the player’s performance score in each subgame. The player’s metacognitive skill score is calculated as a percentage value of his/her performance across all subgames that target the same metacognitive skill. The learning analytics dashboards presented in this paper utilize two different charts: (1) line graphs, which visualize metacognitive skill scores; and (2) scatter plots, which visualize performance scores. The dashboard also features two tabs: (1) a “Brain” tab, which visualizes data about each metacognitive skill; and (2) a “Game” tab, which visualizes data about subgames. The player can filter the data displayed by toggling on and off metacognitive skills or subgames; and changing the date range displayed (see Fig. 1). Figure 1. Line graph of metacognitive skill improvement with Associative Reasoning exploded In the following paragraphs, different visualizations of the dashboard as well as their use cases and benefits for players are explained in more detail. 229 2018 IEEE 18th International Conference on Advanced Learning Technologies 2161-377X/18/$31.00 ©2018 IEEE DOI 10.1109/ICALT.2018.00121