|
|
--- |
|
|
language: |
|
|
- en |
|
|
tags: |
|
|
- sports |
|
|
- pl |
|
|
size_categories: |
|
|
- 10K<n<100K |
|
|
--- |
|
|
# 🏟️ EPL Match Statistics (2000–2024) – Exploratory Data Analysis |
|
|
|
|
|
### Author: Ori Berger |
|
|
### Dataset: [Premier League Match Data (Hugging Face)](https://huggingface.co/datasets/Orib24/Epl_Stats_EDA/blob/main/epl_final.csv) |
|
|
|
|
|
--- |
|
|
|
|
|
## 📊 Overview |
|
|
This project performs an **Exploratory Data Analysis (EDA)** on **9,380 English Premier League matches** from the 2000/01 to 2024/25 seasons. |
|
|
The dataset includes match-level statistics such as goals, shots, corners, fouls, and cards for both home and away teams. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🎯 Objective |
|
|
To analyze which in-game statistics most strongly influence **match outcomes** (`FullTimeResult` = H/D/A) |
|
|
and to identify performance patterns that explain **winning behavior** in football. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧹 Data Cleaning |
|
|
- Verified that no duplicate matches exist. |
|
|
- Checked for missing values (minimal and left unchanged). |
|
|
- Parsed match dates and standardized team names (e.g., “Man Utd” → “Manchester United”). |
|
|
- Converted result columns (`FullTimeResult`, `HalfTimeResult`) to categorical data types. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🚨 Outlier Handling |
|
|
- Outliers found in **shots**, **corners**, and **fouls** were analyzed using z-scores (|z| ≥ 3). |
|
|
- These values were **kept** since they represent authentic extreme matches (e.g., red cards or large wins). |
|
|
|
|
|
--- |
|
|
|
|
|
## 📈 Descriptive Statistics |
|
|
- **Average goals per match:** 2.72 |
|
|
- **Average home goals:** 1.57 | **Average away goals:** 1.15 |
|
|
- **Home advantage:** 46% wins, 25% draws, 29% losses. |
|
|
- **Strong correlation:** goals ↔ shots on target (`r ≈ 0.78`). |
|
|
|
|
|
--- |
|
|
|
|
|
## 🔍 Research Questions & Key Insights |
|
|
1️⃣ **Does playing at home significantly affect match outcomes?** |
|
|
→ Yes — home teams win nearly half their games, confirming a clear home advantage. |
|
|
|
|
|
2️⃣ **How are shots and shots on target related to goals?** |
|
|
→ Strong positive correlation — more shots on target strongly increase goal likelihood. |
|
|
|
|
|
3️⃣ **Do corners reflect attacking dominance?** |
|
|
→ Winning teams average ~2.5 more corners than losing teams. |
|
|
|
|
|
4️⃣ **Do yellow cards or fouls influence match results?** |
|
|
→ Losing teams receive slightly more yellow cards on average, but correlation is weak. |
|
|
|
|
|
5️⃣ **How do goal trends evolve over time?** |
|
|
→ Average goals per match remain steady (~2.7) across the last two decades. |
|
|
|
|
|
--- |
|
|
|
|
|
## 📊 Visualizations |
|
|
- **Histogram:** Distribution of total goals per match. |
|
|
- **Scatter plot:** Shots on target vs goals scored. |
|
|
- **Bar charts:** Averages by match result (shots, corners, cards). |
|
|
- **Line plot:** Average goals per season (2000–2024). |
|
|
|
|
|
Each plot is clearly labeled with titles, axes, and legends. |
|
|
|
|
|
 |
|
|
|
|
|
 |
|
|
|
|
|
 |
|
|
|
|
|
 |
|
|
|
|
|
 |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧠 Conclusions |
|
|
- **Home advantage** is a consistent and statistically significant trend. |
|
|
- **Shots on target** are the strongest predictor of winning matches. |
|
|
- **Corners** serve as a reliable proxy for attacking dominance. |
|
|
- **Disciplinary actions (cards)** have limited predictive value. |
|
|
- Overall, the EPL remains a **balanced and high-scoring league** over time. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧩 Files Included |
|
|
- [`epl_final.csv`](https://huggingface.co/datasets/Orib24/Epl_Stats_EDA/blob/main/epl_final.csv) – dataset |
|
|
- `assignment_ori_berger.ipynb` – notebook with full SQL + Python analysis |
|
|
- `README.md` – summary of results and insights |
|
|
- 'Loom Video'- https://www.loom.com/share/0fadb98589fe473b9222205e6db8b8da |
|
|
<div style="position: relative; padding-bottom: 56.25%; height: 0;"><iframe src="https://www.loom.com/embed/0fadb98589fe473b9222205e6db8b8da" frameborder="0" webkitallowfullscreen mozallowfullscreen allowfullscreen style="position: absolute; top: 0; left: 0; width: 100%; height: 100%;"></iframe></div> |
