Návrh a implementace webové aplikace pro tvorbu soupisek fotbalového turnaje
Loading...
Files
Downloads
9
Date issued
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Vysoká škola báňská – Technická univerzita Ostrava
Location
Signature
Abstract
The aim of this bachelor's thesis is to design and implement a web application intended for creating rosters for a football tournament, specifically the Euro Championship. The outcomes of the thesis include tools for generating optimal team rosters, analyzing player data, and simulating tournament results.
The thesis is structured into several key parts. First, it outlines the theoretical and methodological foundations, including topics such as data acquisition, analysis of football player statistics, roster generation, and tournament simulation using Elo ratings and Poisson regression. It also specifies the functional and non-functional requirements of the resulting application.
The thesis further describes the practical implementation of the data collection process for players and historical matches, as well as the algorithm for automated generation of optimal team rosters considering various formations and maximizing overall team strength. The implementation of the tournament simulation is examined in detail, including the training of a goal prediction model using Scikit-learn, the integration of the Elo rating system, and the use of Monte Carlo simulation with 10,000 iterations to estimate the teams’ success probabilities.
All developed components are integrated into a web application built with the Flask framework. This application allows users to browse generated rosters, view detailed simulation results, and use tools for player data analysis, including interactive visualizations created with the Plotly library. The final application successfully fulfills the main objective of the thesis.
Description
Subject(s)
Web Application, Python, Flask, Football Roster Creation, Tournament Simulation, Elo Rating, Poisson Regression, Monte Carlo Simulation, Web Scraping, Scikit-learn