Project Overview

This project is for Machine Learning practice. I will learn how to predict the winner of football matches in the English Premier League (EPL).

Project Steps

• Scrape match data using request, BeautifulSoup, and pandas.
• Clean the data and get it ready for machine learning using pandas.
• Make predictions about who will win a match using scikit-learn.
• Measure error and improve our predictions.
• Add comments on code

Code

The code will be in two files:

• matches.ipynb - a Jupyter notebook that scrapes our data.
• predictions.ipynb - a Jupyter notebook that makes predictions.

Local Setup

Installation

To follow this project, this is needed locally:

• JupyterLab
• Python 3.8+
• Python packages
  • pandas
  • requests
  • BeautifulSoup
  • scikit-learn
  • html5lib

Data

I will be scraping FBref to get the data first.

For the predictions, I will be using the CSV file with all the data scrapped.

Disclaimer

After running the code I found a couple of issues:

• FBref does not kicks me out of the server regarless of fulfilling the 3 seconds minimun rest betweent requests.
• To have a better dataset I manually set each link of the season I wanted to scrape.