Data Analysis in the BIG DATA scope in Basketball

• Autor(a) principal: Alves, Diogo Filipe Pinto
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/21001
• topic: Big Data; Data analysis; K-Means; Computational and mathematical methods; Unstructured data; Big Data; Análise de dados; Métodos computacionais e matemáticos; Dados não estruturados

Nowadays we are witnessing a great growth in the generation, storage and treatment of large amounts of data. These data are generated by different sources, such as the Web, IoT devices, computers software, smartphone apps, sports and so on. Big Data is a term used for large, varied and complex sets of data, with difficulties in storage, analysis and visualization for later processes or results. The process of searching large amounts of data to reveal hidden patterns and correlations is called Big Data data analysis. This information is useful for companies or organizations and with the help of numerical and computational methods, results can be obtained in a short space of time. For this reason, data implementations in Big Data need to be analysed and executed as accurately as possible. With the amount of data generated, data storage is being crucial. The huge increase in data does not stop and data analysis and visualization are adding to the Big Data era with the amount of data generated by computers, social networks, mobile devices, data collection in sports, etc. This research presents an overview of the content, scope, samples, methods, advantages, challenges and concerns of data analysis in Big Data. Basketball is one of the examples, where we will work with the data, applying types of analysis and data visualization, to understand them and in the end, show their results. Using Clustering algorithms and, with criteria defined in the being of the problems, we will have the same information spread over two or more clusters. Important steps, such as the analysis of each of the indicators and, the objective, we determine rule settings for the expected result. In the results demonstration, we verified that the applied clustering algorithm, K-Means, obtained good results comparing with other data. With the completion of this work, we can better understand the scope of Big Data and apply mathematical clustering methods to extract useful information from large amounts of data.