Using regression to predict the pontuation and place of a soccer team in Campeonato Brasileiro de Futebol

About The Project

I am a soccer lover so I decided to start a project using programming to try to predic with regression the final pontuation of the soccer teams that are participating in Campeonato Brasileiro de Futebol Série A, which is a running point championship with 38 rounds and 20 teams and a very hard one with 7 different winners in the past 18 years.

• This project was developed for the subject Introdução ao aprendizado de máquina e à mineração de dados (PO450) offered by the college Unicamp

Built With

Python in Jupyter Notebook

I decided to try different machine learning approaches to compare the results, so i used linear regression, decision tree, random forest and XGBoost. To get the data to train and to test my machine learning models I used web scraping from the site WorldFootball.net.

Features and scores

Data for training and testing

X_train = atrib_norm[:4560] # 2013-2018

X_test = atrib_norm[4560:5320] # 2019

d_train = pontuacao[:4560] # 2013-2018

d_test = pontuacao[4560:5320] # 2019

• The year of 2020 was used as validation data

Results

For the year of 2020 the models predicted the winner of the championship, Flamengo, from the data of the round 32:

The predictions for 2021 using data from round 12: