Nowadays, technology applied in industries has increased day by day. As a result, different ways of collecting and processing data have been investigated. Machine Learning is one of those ways where, in addition to its applicability in all sectors, it has been increasingly explored in different sports. Furthermore, the analysis of data at a visual level helps in the interpretation and understanding of them. These types of procedures always seek to support in the decision-making work done by coaches, managers and scouting. It can be inherent to any sport and handball is obviously included. This specific investigation addresses methods to create advantages for Handball, introducing predictive analytics. The discovery of promising athletes based on collected variables is one of the biggest challenges in this sport and although the data provided by the Federação de Andebol de Portugal are limited, this study demonstrates a 'direction' of how it can be done based on a single variable. In addition to working in the collection and pre-preparation of sports data, examples of visual presentations such as vertical/horizontal bar graphs and maps are exposed. Finally, Machine Learning algorithms with and without default parameters are used to predict if the player is promising. From this perspective, it can be concluded that, based on formation years, models score is slightly better for Support Vector Machines algorithms despite the proximity of the results. It is important to point out that relevant conclusions were also drawn from the graphs.
© Copyright 2023 NOVA Information Management School. Veröffentlicht von Universidade Nova de Lisboa. Alle Rechte vorbehalten.