FIFA 22 Ultimate Team Recommender

Overview

This is project using Python to scrap FIFA player data from an open source website SoFIFA and apply automated non-interacitve scripts to process crawled data, then fit models to player statistics to setup a recommender (With UI) of players in FIFA22 given user inputs as constraints eg. budget, potential, attribute and etc.

Authors:

Chloe Zhang:

• GitHub

Jinghan Xu:

• jh0220x@gmail.com

Minting Fu

• mminting@student.ubc.ca

Muke Wang

• shaodwaaron@gmail.com

Shiyang Zhang

• harryzhang957@gmail.com

Tony Liang:

• chunqingliang@gmail.com

• GitHub

Usage of the project (Instructions)

1. Without using Docker

To replicate this project, clone this GitHub repository, install the dependencies listed below, and run the following command at the command line/terminal from the root directory of this project:

make all

To reset the repo to a clean state, with no intermediate or results files, run the following command at the command line/terminal from the root directory of this project:

make clean

2. Using Docker to run on Jupyter lab

1. Clone this GitHub repository and run the following code in the terminal

git clone https://github.com/FIFA22-UT-Recommender/fifa22-ultimate-team-recommender

2. Run this firstly in your terminal to pull latest docker image

docker pull tonyliang19/fifa22-ultimate-team-recommender

3. Run the following command to run the container based on the latest image

docker run -it --rm -v /$(pwd):/opt/notebooks/ -p 8888:8888 tonyliang19/fifa22-ultimate-team-recommender

4. After the command runs, copy the last link in your terminal similar to the following:

http://127.0.0.1:8888/lab?token=f4eef0c11762e60a7974f3ea3eb352a4913e70755433398b and open it on any browser like Google Chrome or Mozilla Firefox.

Then you should be able to run and explore the project interactively!

How it works

The principle of our recommended system is that: According to what player the user is looking for, find the player with the lowest price.

For example:

Imagine if a user wants to find a substitution of L.messi in his team. After he input the data of L.messi into our system, we would recommend the slightly cheaper players that is most similar to L.messi(in terms of skills)

1 Train our model

1.1 Label players by their value(price)

example:

Player	Price	Label
L.Messi	€69.5M	1
D.Malen	€28.5M	2
E.Ferguson	€3.6M	3

1.2 Using Quadratic Discriminant Analysis (QDA) model to find decision boundary.

2 Predict the label of player that the user want

2.1 Ask user input

• Postion of the player (eg: LB, ST, CAM)
• Age of the player
• Height of the player
• Different Skill of the player

2.2 Using Quadratic Discriminant Analysis (QDA) model to label the player.

3 Recommend a list of players to user

example:

If label of the player that the user want to find is labeled as 2 by the QDA model, we would recommend the nearest player that labeled as 3

Dependencies

• Python 3.9.5 and Python packages:
  • aiohttp=3.8.1
  • awscli=1.25.2
  • beautifulsoup4=4.11.1
  • bs4=0.0.1
  • certifi=2021.10.8
  • chardet=4.0.0
  • charset-normalizer=2.0.12
  • idna==3.3
  • fifa-pack=0.0.3
  • numpy=1.22.3
  • pandas=1.4.2
  • pathlib=1.0.1
  • python-dateutil=2.8.2
  • pytz=2022.1
  • requests=2.27.1
  • six=1.16.0
  • soupsieve=2.3.2.post1
  • urllib3=1.26.9
• R version 4.1.1 and R packages:
  • bookdown=0.25
  • docopt=0.7.1
  • devtools=2.4.3
  • knitr=1.38
  • rfifa=1.0.0
  • rlang=1.0.2
  • tidyverse=1.3.1
  • tinytex=0.38
  • vctrs=0.4.1
• GNU make 4.3

License

The underlying source code used to format and display the content of this project is licensed under the MIT License