A carregar...
Publicação
VisuaLeague III: Visual Analytics of Multiple Games
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 14.87 MB | Adobe PDF |
Autores
Resumo(s)
Digital data available has been growing over the last years and with it, the need to create representative ways to understand and make use of its potential with visualization techniques that can be applied in different purposes. One of these cases are eSports (electronic sports), considered nowadays a sport with high growth expectation, and for which data analyses can have a significant impact. One of the most popular game type practiced in eSports is the Multiplayer Online Battle Arena (MOBA) genre represented by one of the most popular competitive games, League of Legends (LoL), which will be the case study for this thesis. As many traditional sports, there are various events to have in consideration when observing performance of gameplay. In addition to statistics for each game there is relevant information on players’ positions (spatial data), in a specific period in time (temporal data). Specific events in a game, related with objectives, can also be considered, such as purchasing an item, player kills, destroying towers, or complete objectives. Having a way to analyze and visualize this data helps not only programmers and game designers to improve gameplay but also players, coaches and analysts to improve player performance. The objective of this work is to redesign the previous prototype VisuaLeague II, and propose a new version, VisuaLeague III in order to explore techniques to implement analysis for multiple games, team searches and access to professional games’ training sections, scrims. Common problems presented in the analysis with voluminous amount of data, like cluttering and overlapping, are addressed by adding filters to searches, interaction with the visualizations, aggregation of data, and clustering. The developed prototype, VisuaLeague III was evaluated by professional coaches to understand if the searches and visualization techniques implemented are adequate for analysing players’ performance in a competitive environment. The results demonstrate overall positive attitude with particular interest in analysis for custom games and multiple games analysis as those provide visualizations that do not exist in common tools, specially, regarding spatiotemporal data.
Descrição
Tese de mestrado, Informática, Universidade de Lisboa, Faculdade de Ciências, 2020
Palavras-chave
Visualização de dados de jogos dados espacio-temporais visualização de dados agregados visualização analítica League of Legends Teses de mestrado - 2020