likarajo/premier_league

Predictions for the English Premier League season

Premier League

The Premier League, often referred to as the English Premier League or the EPL outside England, is the top level of the English football league system. It is Contested by 20 clubs every season.

Goal

To predict the result of every fixture

Dependencies

• Glob
• Pandas
• Seaborn
• Scikit-learn

pip install -r requirments.txt

Dataset

• The dataset is taken from https://www.football-data.co.uk/englandm.php

  • Contains data for last 10 seasons of English Premier League, along with the data from the current season.

• Details: https://www.football-data.co.uk/notes.txt

• Fixture is taken from : https://fixturedownload.com/results/epl-2019

Import Dataset

• Read all the data files from previous seasons
• Read only required columns
  • Detailed description: https://www.football-data.co.uk/notes.txt

* Date = Match Date (dd/mm/yy)
* HomeTeam = Home Team
* AwayTeam = Away Team
* HTHG = Half Time Home Team Goals
* HTAG = Half Time Away Team Goals
* HTR = Half Time Result (H=Home Win, D=Draw, A=Away Win)
* FTHG = Full Time Home Team Goals
* FTAG = Full Time Away Team Goals
* FTR = Full Time Result (H=Home Win, D=Draw, A=Away Win)
* HS = Home Team Shots
* AS = Away Team Shots
* HST = Home Team Shots on Target
* AST = Away Team Shots on Target
* HC = Home Team Corners
* AC = Away Team Corners
* HF = Home Team Fouls Committed
* AF = Away Team Fouls Committed
* HY = Home Team Yellow Cards
* AY = Away Team Yellow Cards
* HR = Home Team Red Cards
* AR = Away Team Red Cards

• Concatenate the data from the files together
• Import the concatenated data into a data frame

Obtain team names

• Read the current season's data file
• Obtain the list of home teams from the records
• Find the unique teams from the home teams list

---

Data Analysis

data_analysis.ipynb

Home Team Win Rate

• Total matches: 7430
• Total Home team wins: 3449
• Home wins rate: 46.42 %

Team Statistics

• Games
• Wins
• Win Rate
• Losses
• Draws
• Goals scored
• Scoring rate
• Goals conceded
• Shots
• Shots on Target
• Shots against
• Fouls
• Yellow Cards
• Red Cards

team_stats.csv

---

Data Preprocessing

Keep records of only the teams in the current season

Prepare features and labels

• Features: all columns except 'Date', 'HTR', 'FTR'

18 Features
HomeTeam
AwayTeam
HTHG
HTAG
FTHG
FTAG
HS
AS
HST
AST
HC
AC
HF
AF
HY
AY
HR
AR

• Labels = 'FTR'

Scale and standardise the features

• Center to the mean and component wise scale to unit variance.

Handle categorical values

• Input data needs to be continous variables that are integers
• Convert to dummy variables

Split data to Training and Test sets

• Test data: 50

Create models

Classifiers:

• Logistic Regression
• Support Vector Classifier
• K-Nearest Neighbors

Train and evaluate the models

• Train the model
• Test based on the F1 score and Accuarcy
  • F1 score considers both the precision and the recall of the test to compute the score
  • The F1 score can be interpreted as a weighted average of the precision and recall
  • F1 score reaches its best value at 1 and worst at 0.
  • Accuracy is the ratio of correct predictions to the total predictions

Use the best model for making predictions

• Set the model
• Train the model with training dataset
• Make predictions
• Predict the probability of results (Away team win, draw, Home team win)

Incorporate the result probabilities into the fixture

• Import the fixture
• Set up dataframe to include the fixture along with the features = Test data
• Preprocess the set
• Make predictions

predictions.csv