yanivohayon1/ucl-2021-22

UCL 2021/22 — UEFA Champions League Player Statistics

Dataset Overview

This dataset contains player statistics from the 2021/22 UEFA Champions League season. It consists of 8 CSV files covering different aspects of player performance, with a total of 3,524 rows across all files. After merging all files, the working dataset contains 751 unique players across 44 columns.

Source: Kaggle — UCL Matches & Players Data 2021/22

Target Variable: position — Forward / Midfielder / Defender / Goalkeeper

---

Main Research Goal

The main goal of this project is to understand what makes a player statistically unique based on their position, and how positional roles influence attacking performance, disciplinary behavior, and overall goal contributions in the 2021/22 UEFA Champions League season.

---

Feature Selection

The dataset consists of 8 separate files, each covering a different aspect of player performance. The key features used in this analysis are:

Attacking: goals, assists, shot attempts, shots on target

Defending: tackles, clearances, blocks, interceptions

Disciplinary: yellow cards, red cards, fouls committed

Distribution: passes, pass accuracy

Goalkeeping: saves, clean sheets, goals conceded

Key Stats: minutes played, matches played, position, club

---

Files

File	Description
attacking.csv	Goals, assists, shot attempts
attempts.csv	Shot attempts breakdown
defending.csv	Tackles, clearances, blocks
disciplinary.csv	Yellow cards, red cards, fouls
distributon.csv	Passes, pass accuracy
goalkeeping.csv	Saves, clean sheets
goals.csv	Goal details
key_stats.csv	General player stats
notebook_1 (1).ipynb	Full EDA Notebook

---

Data Cleaning & Decision Making

Step 1 — Merging: I merged all 8 CSV files into one Master DataFrame using player_name, club, and position as keys. The result was 751 unique players × 44 columns.

Step 2 — Missing Values: I found that goalkeeping columns (saved, cleansheets, etc.) had 92.9% missing values — this is expected since only 53 out of 751 players are goalkeepers. I filled those with 0 for non-goalkeepers, and all other numeric columns were filled with the median per position. Final result: 0 missing values.

Step 3 — Duplicates: No duplicate rows were found.

Step 4 — Scaling & Normalization: I checked the range of each key column and found that columns exist on very different scales (e.g. minutes_played up to 1,230 vs goals up to 15). I chose not to apply normalization at this stage since I am performing EDA only and not building an ML model. The original scales are clear and interpretable as-is.

Step 5 — Outlier Detection: I used the IQR method to detect outliers and visualized them using Box Plots. I found 183 outliers in goals and 176 in assists. I decided to keep all outliers because they represent genuine elite performers like Benzema with 15 goals — not data errors.

---

Research Questions & Visual Insights

Question 1: What is the unique statistical profile of each position?

Forwards dominate all attacking statistics with an average of 1.23 goals per player. Goalkeepers play the most minutes (418 on average). Each position has a clearly distinct statistical profile that reflects its role on the pitch.

---

Question 2: Do more shot attempts lead to more goals, or does accuracy matter more?

Shots on target correlate more strongly with goals (r=0.85) than total attempts (r=0.75). This confirms that accuracy matters more than quantity — a player who shoots less but more accurately scores more goals.

---

Question 3: Does a player's position affect the number of fouls committed?

Midfielders commit the most fouls, but defenders receive the most yellow cards (0.065 average). This suggests that defensive positioning leads to more consequential fouls than midfield ones.

---

Question 4: Which club was the most efficient — most goals per shot attempt?

Villarreal was the most efficient club, converting 15% of their shots into goals. Surprisingly, big clubs like Real Madrid and Bayern did not top the list — efficiency is not always about having the best players.

---

Question 5: Do clubs with more assists necessarily score more goals?

Yes — there is a near-perfect correlation of r=0.98 between assists and goals at the club level. Teamwork is the key to attacking success in the UCL.

---

Question 6: Does a player's position affect their goal contributions (goals + assists)?

Forwards and midfielders together account for 84% of all goal contributions (42.5% and 41.3% respectively). Goalkeepers contribute only 0.2% — exactly as expected.

---

Key Findings Summary

• Forwards average 1.23 goals per player — significantly more than any other position
• Shot accuracy (r=0.85) predicts goals better than total attempts (r=0.75)
• Defenders receive the most yellow cards despite midfielders committing more fouls
• Villarreal was the most efficient club, converting 15% of shots into goals
• Assists and goals have a near-perfect correlation (r=0.98) — teamwork is key
• Forwards and midfielders account for 84% of all goal contributions

---

Final Conclusion

The EDA process successfully told the story of UCL 2021/22 player performance. The analysis revealed that position is the strongest predictor of a player's statistical profile — forwards and midfielders dominate goal contributions, while defenders and goalkeepers play a completely different role. Efficiency matters more than volume — Villarreal proved that converting chances well is more important than creating many of them. Finally, teamwork is the true engine of attacking success, as clubs with more assists almost always score more goals.

---

Challenges & Lessons Learned

Challenges: I was on military reserve duty during semester A and did not study Python properly. My main challenge was learning how to write code correctly before and during the project, with the help of Claude AI. Additionally, merging 8 files with different columns caused recurring errors that took time to resolve.

Lessons Learned: At the beginning I did not work in an organized and structured way, which made things increasingly difficult as I progressed — to the point where I had to start over. The most important lesson I learned: working in an organized and structured way, step by step, is the foundation of success.

---

📂 Project Files & Deliverables

File	Description	Link
attacking.csv	Attacking stats	View File
attempts.csv	Shot attempts	View File
defending.csv	Defensive stats	View File
disciplinary.csv	Cards & fouls	View File
distributon.csv	Pass stats	View File
goalkeeping.csv	Goalkeeper stats	View File
goals.csv	Goal details	View File
key_stats.csv	General stats	View File
notebook_1 (1).ipynb	Full EDA Notebook	View Notebook