Publicação
Esports – Video Game Data Analysis
| Resumo: | The phenomenon of Esports has been growing and, with it, the interest in online video games by players and spectators. With technological evolution, it has become increasingly easier to use data collection techniques about the events that take place during a match, generating large volumes of data that can be used to analyze the performance of players and teams. This analysis is of great importance in both personal and professional contexts. Casual players look for methods to understand what mistakes they are making and the optimal way to play certain characters, while in a professional context, the focus is mostly on understanding what strategies are used by other teams and how to counter them. For the analysis of this volume of data to be effective, it is fundamental to explore data analysis mechanisms combined with visualization techniques (visual analytics) applied to spatio-temporal data and the various types of events during a match that are of interest to players, coaches, and analysts. These events can range from a player’s position (space) at a given instant (time) to more game-specific events, such as where the player died. The goal of this project is to explore and apply machine learning algorithms to spatio-temporal data to discover patterns in player behaviors while continuing the work on visual analytics of video game data. The investigation extends to exploring datasets from new games, ultimately leading to the selection of League of Legends (LoL) as the focal point for in-depth analysis. One significant challenge in this pursuit is the scarcity of readily available datasets featuring spatio-temporal data. To overcome this obstacle, the research project involves the creation of spatio-temporal datasets from the selected game, LoL, through data collection facilitated by the Riot API. In summary, this research project not only builds on previous work but also introduces new data analysis techniques, notably the clustering of spatio-temporal data, to uncover possibly hidden patterns of player behaviors in the world of League of Legends. The results obtained provide valuable insights on the players, particularly focusing on the jungler role. They provide information regarding potential death patterns and the most frequently visited locations on the map as the game progresses. Additionally, these results make it possible to observe differences in spatio-temporal data across various game patches. The culmination of these efforts promises valuable insights into the gaming ecosystem, with potential applications in game design, player engagement, and beyond. |
|---|---|
| Autores principais: | Roque, Rui Pedro Fernandes |
| Assunto: | Esports Visualização analítica Dados espácio-temporais Algoritmos de aprendizagem automática Análise de dados Teses de mestrado - 2023 |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | The phenomenon of Esports has been growing and, with it, the interest in online video games by players and spectators. With technological evolution, it has become increasingly easier to use data collection techniques about the events that take place during a match, generating large volumes of data that can be used to analyze the performance of players and teams. This analysis is of great importance in both personal and professional contexts. Casual players look for methods to understand what mistakes they are making and the optimal way to play certain characters, while in a professional context, the focus is mostly on understanding what strategies are used by other teams and how to counter them. For the analysis of this volume of data to be effective, it is fundamental to explore data analysis mechanisms combined with visualization techniques (visual analytics) applied to spatio-temporal data and the various types of events during a match that are of interest to players, coaches, and analysts. These events can range from a player’s position (space) at a given instant (time) to more game-specific events, such as where the player died. The goal of this project is to explore and apply machine learning algorithms to spatio-temporal data to discover patterns in player behaviors while continuing the work on visual analytics of video game data. The investigation extends to exploring datasets from new games, ultimately leading to the selection of League of Legends (LoL) as the focal point for in-depth analysis. One significant challenge in this pursuit is the scarcity of readily available datasets featuring spatio-temporal data. To overcome this obstacle, the research project involves the creation of spatio-temporal datasets from the selected game, LoL, through data collection facilitated by the Riot API. In summary, this research project not only builds on previous work but also introduces new data analysis techniques, notably the clustering of spatio-temporal data, to uncover possibly hidden patterns of player behaviors in the world of League of Legends. The results obtained provide valuable insights on the players, particularly focusing on the jungler role. They provide information regarding potential death patterns and the most frequently visited locations on the map as the game progresses. Additionally, these results make it possible to observe differences in spatio-temporal data across various game patches. The culmination of these efforts promises valuable insights into the gaming ecosystem, with potential applications in game design, player engagement, and beyond. |
|---|