Soccer Players Data Analyst and Similar Players Finder

Clone the repo

git clone https://github.com/mchien15/datascience.git

then navigate to repo's folder

Install required packages

pip install -r requirements.txt

This application also requires Docker, so if you haven't already installed it on your computer, please follow this INTRUCTION

Prepare the data

Start our data lake infrastructure

docker compose -f docker-compose.yml up -d

Clean data (drop replical columns, rename columns, convert csv files to parquet)

python clean_data.py

Generate data and push them to MinIO

python utils/export_data_to_datalake.py

Create data schema

After pushing your files to MinIO, please run the following command to execute the trino container:

docker exec -ti datalake-trino bash

When you are already inside the trino container, run trino to enter the interactive mode

After that, copy and run this chunk of commands to register a new schema for the data:

CREATE SCHEMA IF NOT EXISTS datalake.data_big_5_leagues WITH (location = 's3://data-big-5-leagues/');

CREATE TABLE IF NOT EXISTS datalake.data_big_5_leagues.all_leagues (
    Date VARCHAR,
    Name VARCHAR,
    Round VARCHAR,
    Venue VARCHAR,
    Result VARCHAR,
    Squad VARCHAR,
    Opponent VARCHAR,
    Start VARCHAR,
    Pos VARCHAR,
    Min DOUBLE,
    Cmp DOUBLE,
    PassAtt DOUBLE,
    CmpPct DOUBLE,
    PassTotDist DOUBLE,
    PassPrgDist DOUBLE,
    Cmp1 DOUBLE,
    Att1 DOUBLE,
    CmpPct1 DOUBLE,
    Cmp2 DOUBLE,
    Att2 DOUBLE,
    CmpPct2 DOUBLE,
    Cmp3 DOUBLE,
    Att3 DOUBLE,
    CmpPct3 DOUBLE,
    Ast DOUBLE,
    xAG DOUBLE,
    xA DOUBLE,
    KP DOUBLE,
    PassFinThird DOUBLE,
    PPA DOUBLE,
    CrsPA DOUBLE,
    PrgP DOUBLE,
    ID VARCHAR,
    SCA DOUBLE,
    PassLiveShot DOUBLE,
    PassDeadShot DOUBLE,
    TO DOUBLE,
    ShLSh DOUBLE,
    Fld DOUBLE,
    DefShot DOUBLE,
    GCA DOUBLE,
    PassLiveGoal DOUBLE,
    PassDeadGoal DOUBLE,
    TO1 DOUBLE,
    ShGoal DOUBLE,
    FldGoal DOUBLE,
    DefGoal DOUBLE,
    Tkl DOUBLE,
    TklW DOUBLE,
    TacklesDef3rd DOUBLE,
    TacklesMid3rd DOUBLE,
    TacklesAtt3rd DOUBLE,
    DribTackled DOUBLE,
    DribContest DOUBLE,
    DribTackledPct DOUBLE,
    Lost DOUBLE,
    Blocks DOUBLE,
    BlockSh DOUBLE,
    Pass DOUBLE,
    Int DOUBLE,
    TklPlusInt DOUBLE,
    Clr DOUBLE,
    Err DOUBLE,
    Touches DOUBLE,
    DefPen DOUBLE,
    TouchDef3rd DOUBLE,
    TouchMid3rd DOUBLE,
    TouchAtt3rd DOUBLE,
    AttPen DOUBLE,
    Live DOUBLE,
    Att DOUBLE,
    Succ DOUBLE,
    SuccPct DOUBLE,
    Tkld DOUBLE,
    TkldPct DOUBLE,
    Carries DOUBLE,
    TotDist DOUBLE,
    PrgDist DOUBLE,
    PrgC DOUBLE,
    CarriesFinThird DOUBLE,
    CPA DOUBLE,
    Mis DOUBLE,
    Dis DOUBLE,
    Rec DOUBLE,
    PrgR DOUBLE,
    Gls DOUBLE,
    PK DOUBLE,
    PKatt DOUBLE,
    Sh DOUBLE,
    SoT DOUBLE,
    CrdY DOUBLE,
    CrdR DOUBLE,
    xG DOUBLE,
    npxG DOUBLE
) WITH (
    external_location = 's3://data-big-5-leagues/players/',
    format = 'PARQUET'
);

Run the Streamlit app

Open the new terminal or run exit twice, then run this command

streamlit run Main_Page.py

Visit the URL displayed in the terminal (usually http://localhost:8501) to interact with the app

Some of the features of the app

Scouting Report and Similar Players Finder

Radar chart for players comparison

For each position, there will be different stats to be used to compare the players. For examples, these are the plots for comparing Messi - Neymar and Thiago Silva - Van Dijk

Scatter plot for metrics comparison