hbbhsy/european-soccer-eda

European-Soccer-EDA

Data

This dataset contains 25k+ matches, players & teams attributes for European Professional Football, including:

• +25,000 matches
  
• +10,000 players
  
• 11 European Countries with their lead championship
  
• Seasons 2008 to 2016
  
• Players and Teams' attributes* sourced from EA Sports' FIFA video game series, including the weekly updates
  
• Team line up with squad formation (X, Y coordinates)
  
• Betting odds from up to 10 providers
  
• Detailed match events (goal types, possession, corner, cross, fouls, cards etc...) for +10,000 matches
  

Source: Kaggle

https://www.kaggle.com/hugomathien/soccer

Understanding the data

I begin my EDA by looking at my datasets and try to understand my datasets. My data contains a Country table and a League table. I combined the two table into one.

Country	League
Belgium	Belgium Jupiler League
England	England Premier League
France	France Ligue 1
Germany	Germany 1. Bundesliga
Italy	Italy Serie A
Netherlands	Netherlands Eredivisie
Poland	Poland Ekstraklasa
Portugal	Portugal Liga ZON Sagres
Scotland	Scotland Premier League
Spain	Spain LIGA BBVA
Switzerland	Switzerland Super League

The match table contains information of every match events from season 2008/2009 to 2015/2016.

Teams

From the team attribute table, I'm able to look at the correlations between each features and plot the heatmap. The top stronge correlations are shown between the defence pressure and defence team width, defence pressure and defence aggression build up play passing and build up play speed.

Therefore we may say that in order to form a strong defence, a team may wants to design a wider back field squad format.

Players

Same as the team table, I started to look at the players by plot the heatmap.

One interesting observation from this heatmap is, for a player, the height and weight shows a strong positve correlation with his strength,but shows negative correlations with most of his features such as balance, acceleration, agility, etc. I then plotted the distributions for the player's height and weight.

It seems like the two features are normal distributions.

Does height and weight affect on a player's overall rating? Not obviously.

Matches

Base on the Match table, I computed the winning rates for each team. And I was curious on how the team attributes affect the winning rates.

Using the winning rate, I ranked the top ten and bottom ten teams:

Top Ten Teams

Ranking	Team
1	FC Barcelona
2	Real Madrid CF
3	SL Benfica
4	FC Porto
5	Celtic
6	Rangers
7	FC Bayern Munich
8	Ajax
9	PSV
10	RSC Anderlecht

Bottom Ten Teams

Ranking	Team
296	SpVgg Greuther Fürth
295	Córdoba CF
294	AC Arles-Avignon
293	DSC Arminia Bielefeld
292	FC Dordrecht
291	Dunfermline Athletic
290	ES Troyes AC
289	FC Penafiel
288	Pescara
287	Reading

Hypothesis

As a non soccer fan, I always hear people say that home teams has advantages over away teams. I wonder if that's true for soccer. For each team, I obtained the winning rates when it played as home team and away time.

Set my null hypothesis:

I used Welsh's T-test on home winning rates and away winnning rates.

T-statistic	p-value
14.2052	1.832e-39

The p-value is small enough to reject the H₀.

$$\mu$$