DATA 602 - FINAL PROJECT PROPOSAL

Author

Pascal Hermann Kouogang Tafo

CONTEXT

As an aspired sport analyst scientist, Soccer or football has always been my passion therefore i thought it will be a great idea to base my final project on this topic and especially on the evolution of soccer in what is considered the best league in the world nowadays.

1) RESEARCH QUESTION

A way to improve performance in a club resides in the signing of new players whom the club believe will bring more competitive environment. In the Premier league, the amount in players’ transfers has increased exponentially the last recent years. We believe it is almost legitimate to ask ourselves these following questions: - Has money influenced the performance andevolution of English Premier League teams over the last 10 seasons? - In other terms, is the positive relationship between the clubs that spent the most money in transfer and their final position in the season standing? To respond at this question, we will measure the net spending of the clubs (ration between transfer fee spent and received) and their performance by the points and their final league position at the end of each season.

2) Justification : Why is this relevant to me ?

This is relevant to me because as a soccer fan, I am worried about the trajectory observed lately where we see a gap that is been created in teams budget forcing some owners to take urge financial risks that leads in some cases in the disappearances of the historical clubs. Moreover, soccer has become more than a sport but a major global business. Therefore, analyzing historical financial and performance data, clubs can evaluate whether their transfer investments are associated with better results and identify opportunities to improve recruitment efficiency.

3) Data Sources :

Because my research rely on finding the relationship on the players’ transfers and their performances, I will be collecting my data from publicly available online datasets combining historical player transfer records with premier league season-end team statistics.

- Dataset 1: Premier league Players Transfers: The dataset is composed of historical player transfer information for Premier League clubs. Direct Csv file: https://raw.githubusercontent.com/ewenme/transfers/master/data/premier-league.csv

- Dataset 2: Premier League Team-Season Records This dataset contains season-level statistics for Premier League clubs. Direct CSV file: https://raw.githubusercontent.com/thomjjames/premier-league-team-records/main/datasets/team_season_records.csv

4) LIBRARIES potential being used

The following python libraries will be used to read, clean, analyze,visualize and predict.

  • Pandas library should help me read both datasets which are delivered as CSV files. As we know, Pandas will be used to import the data into DataFrames, clean missing values, standardize club names, aggregate transfer fees, merge datasets, and calculate summary statistics.

  • Numpy library to facilitate mathematical operations on arrays, handling missing numerical values, and when calculating net spend, points per million and potentially transformations of skewed sending values.

  • Matplotlib and Seaborn libraries will be used for the EDA visualizations: histograms,scatter plots and bar plots. the goal here is show the changes in transfer expenditure and team performance over the 10-year period.

  • SciPy and Scikit-learn libraries: SciPy will provide rank correlations. Scikit-learn may be used totrain and test models that predict points from spending.

5) EDA and summary statistics.**

The exploratory data analysis will investigate how transfer spending and Premier League teams performance have evolved over the selected 10 seasons. It is important to know that the premier league is composed of 20 clubs which means we will be basing our analysis over 200 club-seasons (loans excluded, missing fees counted as zero). The first step will be to inspect the datasets, identify missing values, check for duplicate records, standardize club names, and ensure that transfer fees and season identifiers are consistent. Transfer records will then be aggregated into club-season totals before being merged with the team performance dataset. For the summary statistic, The following statistics will be calculated: Mean, Median, Standard Deviation, Minimum, Maximum, Quartiles. For EDA, data visualization such as histograms, scatter plots and bar plots would seem reasonable to capture information about each club and season. I plan to look at spending against points for each season, total spending by club, league-wide spending over time,and points per million euros. As a later step, I will train and test simple models that predict points from spending.

Below, is the first glance of the code where we load the data, check the missing data and give us an overview of the calculation of the summary statistics.

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Load the prepared club-season dataset
df = pd.read_csv("premier_league_financial_performance.csv")

# Inspect the dataset
print(df.head())
print(df.info())

# Check missing values and duplicate records
print(df.isnull().sum())
print("Duplicate rows:", df.duplicated().sum())

# Calculate summary statistics

columns = [
    "transfer_expenditure",
    "transfer_income",
    "net_spending",
    "points",
    "position",
    "goal_difference",
    "wins"
]

print(df[columns].describe().round(2))

# 1. Total transfer expenditure by season
season_spending = (
    df.groupby("season")["transfer_expenditure"]
    .sum()
    .reset_index()
)