Publicação
Bringing objectivity and predictability to one of the most diverse and opiniated sports in the world by leveraging data
| dc.contributor.advisor | Jesus, Frederico Miguel Campos Cruz Ribeiro de | |
| dc.contributor.author | Sousa, David Silva de | |
| dc.date.accessioned | 2022-07-27T14:05:54Z | |
| dc.date.available | 2022-07-27T14:05:54Z | |
| dc.date.issued | 2022-05-20 | |
| dc.description | Internship Report presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science | pt_PT |
| dc.description.abstract | Football institutions look to gain a competitive edge on and off the pitch. In this sense, this report showcases two data-led performance analyses that provide objective and measureable insights to support decision-making within clubs. Firstly, Pass Sonar, an interactive tool that represents the frequency and distance of players' passes by direction in order to understand historical passing behaviors. Secondly, Valuing Actions by Estimating Probabilities (VAEP), a data-driven framework for assessing the value of individual technical actions performed by players during football matches. This framework values all offensive and defensive action types (e.g., passes, shots, clearances, and tackles) using event data whereas lacking teams' positional structure, environment around the ball, and disregarding tactical and physical contributions captured by tracking systems. On the other hand, this method makes it scalable for teams to use as tracking systems can be expensive. Notwithstanding, both data-centric solutions focus on player-level and team-level performance evaluation relative to technical aspects of the game across the English Premier League during 2019/20 season. Thus, characterizing teams’ playing styles using Latent Dirichlet Allocation (LDA) algorithm includes a tactical element in future work. | pt_PT |
| dc.identifier.tid | 203044614 | pt_PT |
| dc.identifier.uri | http://hdl.handle.net/10362/142482 | |
| dc.language.iso | eng | pt_PT |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | pt_PT |
| dc.subject | Sports Analytics | pt_PT |
| dc.subject | Data Science | pt_PT |
| dc.subject | Probabilistic Classification | pt_PT |
| dc.subject | Valuing Actions | pt_PT |
| dc.subject | Representing Passes | pt_PT |
| dc.subject | Playing Styles | pt_PT |
| dc.subject | Event Data | pt_PT |
| dc.subject | English Premier League | pt_PT |
| dc.title | Bringing objectivity and predictability to one of the most diverse and opiniated sports in the world by leveraging data | pt_PT |
| dc.type | master thesis | |
| dspace.entity.type | Publication | |
| rcaap.rights | openAccess | pt_PT |
| rcaap.type | masterThesis | pt_PT |
| thesis.degree.name | Mestrado em Ciência de Dados e Métodos Analíticos Avançados, especialização em Ciência de Dados | pt_PT |