Visual analytics and team strategies in online games

Abstract

The eSports (electronic sports) phenomenon has been growing and so does the interest in online video games, from players and spectators. With technological advancements it has become easier to use techniques to retrieve data about the events occurring during a game, generating big volumes of data that can be used for a performance analysis. Casual players are looking for methods to better themselves overall or with specific characters, whereas, in a professional context, the focus is to study other teams and how to defeat them. For efficiency, it is imperative to explore data analysis mechanisms combined with visualisation techniques (visual analytics) applied to spatiotemporal data and to various relevant events during a match such as a player’s position (space) in a given instant (time) or, for example, the position where the player died. The goal of this project is the study of previous work and the development and ap plication of the acquired knowledge in analytic visualisation techniques to League of Legends[31] (LoL) spatiotemporal datasets. The developed tool used Tableau Desktop[24] to create a series of dashboards depicting the behaviour of multiple LoL matches, using the Riot API (Application Programming Interface) provided dataset, and clustering algorithms. The tool was evaluated by a team of semi-professional players in order to understand if the visualisation techniques and data used was adequate, useful or innovative compared to already existing tools for game analysis and the players’ needs. The results were mostly positive, with the participants pointing out the interactivity of the visualisations and ability of analysing multiple games as an advantage compared to existing tools. To conclude, even though spatiotemporal data is not yet implemented in MOBA (Multiplayer Online Battle Arena) videogame analysis tools, it is still relevant for the players’ personal goals and overall an interesting approach.