Practical Machine Learning with R and Python – Part 2


In this 2nd part of the series “Practical Machine Learning with R and Python – Part 2”, I continue where I left off in my first post Practical Machine Learning with R and Python – Part 2. In this post I cover the some classification algorithmns and cross validation. Specifically I touch
-Logistic Regression
-K Nearest Neighbors (KNN) classification
-Leave out one Cross Validation (LOOCV)
-K Fold Cross Validation
in both R and Python.

As in my initial post the algorithms are based on the following courses.

You can download this R Markdown file along with the data from Github. I hope these posts can be used as a quick reference in R and Python and Machine Learning.I have tried to include the coolest part of either course in this post.

The following classification problem is based on Logistic Regression. The data is an included data set in Scikit-Learn, which I have saved as csv and use it also for R. The fit of a classification Machine Learning Model depends on how correctly classifies the data. There are several measures of testing a model’s classification performance. They are

Accuracy = TP + TN / (TP + TN + FP + FN) – Fraction of all classes correctly classified
Precision = TP / (TP + FP) – Fraction of correctly classified positives among those classified as positive
Recall = TP / (TP + FN) Also known as sensitivity, or True Positive Rate (True positive) – Fraction of correctly classified as positive among all positives in the data
F1 = 2 * Precision * Recall / (Precision + Recall)

1a. Logistic Regression – R code

The caret and e1071 package is required for using the confusionMatrix call

source("RFunctions.R")
library(dplyr)
library(caret)
library(e1071)
# Read the data (from sklearn)
cancer <- read.csv("cancer.csv")
# Rename the target variable
names(cancer) <- c(seq(1,30),"output")
# Split as training and test sets
train_idx <- trainTestSplit(cancer,trainPercent=75,seed=5)
train <- cancer[train_idx, ]
test <- cancer[-train_idx, ]

# Fit a generalized linear logistic model, 
fit=glm(output~.,family=binomial,data=train,control = list(maxit = 50))
# Predict the output from the model
a=predict(fit,newdata=train,type="response")
# Set response >0.5 as 1 and <=0.5 as 0
b=ifelse(a>0.5,1,0)
# Compute the confusion matrix for training data
confusionMatrix(b,train$output)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction   0   1
##          0 154   0
##          1   0 272
##                                      
##                Accuracy : 1          
##                  95% CI : (0.9914, 1)
##     No Information Rate : 0.6385     
##     P-Value [Acc > NIR] : < 2.2e-16  
##                                      
##                   Kappa : 1          
##  Mcnemar's Test P-Value : NA         
##                                      
##             Sensitivity : 1.0000     
##             Specificity : 1.0000     
##          Pos Pred Value : 1.0000     
##          Neg Pred Value : 1.0000     
##              Prevalence : 0.3615     
##          Detection Rate : 0.3615     
##    Detection Prevalence : 0.3615     
##       Balanced Accuracy : 1.0000     
##                                      
##        'Positive' Class : 0          
## 
m=predict(fit,newdata=test,type="response")
n=ifelse(m>0.5,1,0)
# Compute the confusion matrix for test output
confusionMatrix(n,test$output)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction  0  1
##          0 52  4
##          1  5 81
##                                           
##                Accuracy : 0.9366          
##                  95% CI : (0.8831, 0.9706)
##     No Information Rate : 0.5986          
##     P-Value [Acc > NIR] : <2e-16          
##                                           
##                   Kappa : 0.8677          
##  Mcnemar's Test P-Value : 1               
##                                           
##             Sensitivity : 0.9123          
##             Specificity : 0.9529          
##          Pos Pred Value : 0.9286          
##          Neg Pred Value : 0.9419          
##              Prevalence : 0.4014          
##          Detection Rate : 0.3662          
##    Detection Prevalence : 0.3944          
##       Balanced Accuracy : 0.9326          
##                                           
##        'Positive' Class : 0               
## 

1b. Logistic Regression – Python code

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
os.chdir("C:\\Users\\Ganesh\\RandPython")
from sklearn.datasets import make_classification, make_blobs

from sklearn.metrics import confusion_matrix
from matplotlib.colors import ListedColormap
from sklearn.datasets import load_breast_cancer
# Load the cancer data
(X_cancer, y_cancer) = load_breast_cancer(return_X_y = True)
X_train, X_test, y_train, y_test = train_test_split(X_cancer, y_cancer,
                                                   random_state = 0)
# Call the Logisitic Regression function
clf = LogisticRegression().fit(X_train, y_train)
fig, subaxes = plt.subplots(1, 1, figsize=(7, 5))
# Fit a model
clf = LogisticRegression().fit(X_train, y_train)

# Compute and print the Accuray scores
print('Accuracy of Logistic regression classifier on training set: {:.2f}'
     .format(clf.score(X_train, y_train)))
print('Accuracy of Logistic regression classifier on test set: {:.2f}'
     .format(clf.score(X_test, y_test)))
y_predicted=clf.predict(X_test)
# Compute and print confusion matrix
confusion = confusion_matrix(y_test, y_predicted)
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
print('Accuracy: {:.2f}'.format(accuracy_score(y_test, y_predicted)))
print('Precision: {:.2f}'.format(precision_score(y_test, y_predicted)))
print('Recall: {:.2f}'.format(recall_score(y_test, y_predicted)))
print('F1: {:.2f}'.format(f1_score(y_test, y_predicted)))
## Accuracy of Logistic regression classifier on training set: 0.96
## Accuracy of Logistic regression classifier on test set: 0.96
## Accuracy: 0.96
## Precision: 0.99
## Recall: 0.94
## F1: 0.97

2. Dummy variables

The following R and Python code show how dummy variables are handled in R and Python. Dummy variables are categorival variables which have to be converted into appropriate values before using them in Machine Learning Model For e.g. if we had currency as ‘dollar’, ‘rupee’ and ‘yen’ then the dummy variable will convert this as
dollar 0 0 0
rupee 0 0 1
yen 0 1 0

2a. Logistic Regression with dummy variables- R code

# Load the dummies library
library(dummies) 
df <- read.csv("adult1.csv",stringsAsFactors = FALSE,na.strings = c(""," "," ?"))

# Remove rows which have NA
df1 <- df[complete.cases(df),]
dim(df1)
## [1] 30161    16
# Select specific columns
adult <- df1 %>% dplyr::select(age,occupation,education,educationNum,capitalGain,
                               capital.loss,hours.per.week,native.country,salary)
# Set the dummy data with appropriate values
adult1 <- dummy.data.frame(adult, sep = ".")

#Split as training and test
train_idx <- trainTestSplit(adult1,trainPercent=75,seed=1111)
train <- adult1[train_idx, ]
test <- adult1[-train_idx, ]

# Fit a binomial logistic regression
fit=glm(salary~.,family=binomial,data=train)
# Predict response
a=predict(fit,newdata=train,type="response")
# If response >0.5 then it is a 1 and 0 otherwise
b=ifelse(a>0.5,1,0)
confusionMatrix(b,train$salary)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction     0     1
##          0 16065  3145
##          1   968  2442
##                                           
##                Accuracy : 0.8182          
##                  95% CI : (0.8131, 0.8232)
##     No Information Rate : 0.753           
##     P-Value [Acc > NIR] : < 2.2e-16       
##                                           
##                   Kappa : 0.4375          
##  Mcnemar's Test P-Value : < 2.2e-16       
##                                           
##             Sensitivity : 0.9432          
##             Specificity : 0.4371          
##          Pos Pred Value : 0.8363          
##          Neg Pred Value : 0.7161          
##              Prevalence : 0.7530          
##          Detection Rate : 0.7102          
##    Detection Prevalence : 0.8492          
##       Balanced Accuracy : 0.6901          
##                                           
##        'Positive' Class : 0               
## 
# Compute and display confusion matrix
m=predict(fit,newdata=test,type="response")
## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading
n=ifelse(m>0.5,1,0)
confusionMatrix(n,test$salary)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction    0    1
##          0 5263 1099
##          1  357  822
##                                           
##                Accuracy : 0.8069          
##                  95% CI : (0.7978, 0.8158)
##     No Information Rate : 0.7453          
##     P-Value [Acc > NIR] : < 2.2e-16       
##                                           
##                   Kappa : 0.4174          
##  Mcnemar's Test P-Value : < 2.2e-16       
##                                           
##             Sensitivity : 0.9365          
##             Specificity : 0.4279          
##          Pos Pred Value : 0.8273          
##          Neg Pred Value : 0.6972          
##              Prevalence : 0.7453          
##          Detection Rate : 0.6979          
##    Detection Prevalence : 0.8437          
##       Balanced Accuracy : 0.6822          
##                                           
##        'Positive' Class : 0               
## 

2b. Logistic Regression with dummy variables- Python code

Pandas has a get_dummies function for handling dummies

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
# Read data
df =pd.read_csv("adult1.csv",encoding="ISO-8859-1",na_values=[""," "," ?"])
# Drop rows with NA
df1=df.dropna()
print(df1.shape)
# Select specific columns
adult = df1[['age','occupation','education','educationNum','capitalGain','capital-loss', 
             'hours-per-week','native-country','salary']]

X=adult[['age','occupation','education','educationNum','capitalGain','capital-loss', 
             'hours-per-week','native-country']]
# Set approporiate values for dummy variables
X_adult=pd.get_dummies(X,columns=['occupation','education','native-country'])
y=adult['salary']

X_adult_train, X_adult_test, y_train, y_test = train_test_split(X_adult, y,
                                                   random_state = 0)
clf = LogisticRegression().fit(X_adult_train, y_train)

# Compute and display Accuracy and Confusion matrix
print('Accuracy of Logistic regression classifier on training set: {:.2f}'
     .format(clf.score(X_adult_train, y_train)))
print('Accuracy of Logistic regression classifier on test set: {:.2f}'
     .format(clf.score(X_adult_test, y_test)))
y_predicted=clf.predict(X_adult_test)
confusion = confusion_matrix(y_test, y_predicted)
print('Accuracy: {:.2f}'.format(accuracy_score(y_test, y_predicted)))
print('Precision: {:.2f}'.format(precision_score(y_test, y_predicted)))
print('Recall: {:.2f}'.format(recall_score(y_test, y_predicted)))
print('F1: {:.2f}'.format(f1_score(y_test, y_predicted)))
## (30161, 16)
## Accuracy of Logistic regression classifier on training set: 0.82
## Accuracy of Logistic regression classifier on test set: 0.81
## Accuracy: 0.81
## Precision: 0.68
## Recall: 0.41
## F1: 0.51

3a – K Nearest Neighbors Classification – R code

The Adult data set is taken from UCI Machine Learning Repository

source("RFunctions.R")
df <- read.csv("adult1.csv",stringsAsFactors = FALSE,na.strings = c(""," "," ?"))
# Remove rows which have NA
df1 <- df[complete.cases(df),]
dim(df1)
## [1] 30161    16
# Select specific columns
adult <- df1 %>% dplyr::select(age,occupation,education,educationNum,capitalGain,
                               capital.loss,hours.per.week,native.country,salary)
# Set dummy variables
adult1 <- dummy.data.frame(adult, sep = ".")

#Split train and test as required by KNN classsification model
train_idx <- trainTestSplit(adult1,trainPercent=75,seed=1111)
train <- adult1[train_idx, ]
test <- adult1[-train_idx, ]
train.X <- train[,1:76]
train.y <- train[,77]
test.X <- test[,1:76]
test.y <- test[,77]

# Fit a model for 1,3,5,10 and 15 neighbors
cMat <- NULL
neighbors <-c(1,3,5,10,15)
for(i in seq_along(neighbors)){
    fit =knn(train.X,test.X,train.y,k=i)
    table(fit,test.y)
    a<-confusionMatrix(fit,test.y)
    cMat[i] <- a$overall[1]
    print(a$overall[1])
}
##  Accuracy 
## 0.7835831 
##  Accuracy 
## 0.8162047 
##  Accuracy 
## 0.8089113 
##  Accuracy 
## 0.8209787 
##  Accuracy 
## 0.8184591
#Plot the Accuracy for each of the KNN models
df <- data.frame(neighbors,Accuracy=cMat)
ggplot(df,aes(x=neighbors,y=Accuracy)) + geom_point() +geom_line(color="blue") +
    xlab("Number of neighbors") + ylab("Accuracy") +
    ggtitle("KNN regression - Accuracy vs Number of Neighors (Unnormalized)")

3b – K Nearest Neighbors Classification – Python code

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
from sklearn.neighbors import KNeighborsClassifier
from sklearn.preprocessing import MinMaxScaler

# Read data
df =pd.read_csv("adult1.csv",encoding="ISO-8859-1",na_values=[""," "," ?"])
df1=df.dropna()
print(df1.shape)
# Select specific columns
adult = df1[['age','occupation','education','educationNum','capitalGain','capital-loss', 
             'hours-per-week','native-country','salary']]

X=adult[['age','occupation','education','educationNum','capitalGain','capital-loss', 
             'hours-per-week','native-country']]
             
#Set values for dummy variables
X_adult=pd.get_dummies(X,columns=['occupation','education','native-country'])
y=adult['salary']

X_adult_train, X_adult_test, y_train, y_test = train_test_split(X_adult, y,
                                                   random_state = 0)
                                                   
# KNN classification in Python requires the data to be scaled. 
# Scale the data
scaler = MinMaxScaler()
X_train_scaled = scaler.fit_transform(X_adult_train)
# Apply scaling to test set also
X_test_scaled = scaler.transform(X_adult_test)
# Compute the KNN model for 1,3,5,10 & 15 neighbors
accuracy=[]
neighbors=[1,3,5,10,15]
for i in neighbors:
    knn = KNeighborsClassifier(n_neighbors = i)
    knn.fit(X_train_scaled, y_train)
    accuracy.append(knn.score(X_test_scaled, y_test))
    print('Accuracy test score: {:.3f}'
        .format(knn.score(X_test_scaled, y_test)))

# Plot the models with the Accuracy attained for each of these models    
fig1=plt.plot(neighbors,accuracy)
fig1=plt.title("KNN regression - Accuracy vs Number of neighbors")
fig1=plt.xlabel("Neighbors")
fig1=plt.ylabel("Accuracy")
fig1.figure.savefig('foo1.png', bbox_inches='tight')
## (30161, 16)
## Accuracy test score: 0.749
## Accuracy test score: 0.779
## Accuracy test score: 0.793
## Accuracy test score: 0.804
## Accuracy test score: 0.803

Output image:

4 MPG vs Horsepower

The following scatter plot shows the non-linear relation between mpg and horsepower. This will be used as the data input for computing K Fold Cross Validation Error

4a MPG vs Horsepower scatter plot – R Code

df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))
df2 <- df1 %>% dplyr::select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]
ggplot(df3,aes(x=horsepower,y=mpg)) + geom_point() + xlab("Horsepower") + 
    ylab("Miles Per gallon") + ggtitle("Miles per Gallon vs Hosrsepower")

4b MPG vs Horsepower scatter plot – Python Code

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
autoDF3=autoDF2.dropna()
autoDF3.shape
#X=autoDF3[['cylinder','displacement','horsepower','weight']]
X=autoDF3[['horsepower']]
y=autoDF3['mpg']

fig11=plt.scatter(X,y)
fig11=plt.title("KNN regression - Accuracy vs Number of neighbors")
fig11=plt.xlabel("Neighbors")
fig11=plt.ylabel("Accuracy")
fig11.figure.savefig('foo11.png', bbox_inches='tight')

5 K Fold Cross Validation

K Fold Cross Validation is a technique in which the data set is divided into K Folds or K partitions. The Machine Learning model is trained on K-1 folds and tested on the Kth fold i.e.
we will have K-1 folds for training data and 1 for testing the ML model. Since we can partition this as C_{1}^{K} or K choose 1, there will be K such partitions. The K Fold Cross
Validation estimates the average validation error that we can expect on a new unseen test data.

The formula for K Fold Cross validation is as follows

MSE_{K} = \frac{\sum (y-yhat)^{2}}{n_{K}}
and
n_{K} = \frac{N}{K}
and
CV_{K} = \sum_{K=1}^{K} (\frac{n_{K}}{N}) MSE_{K}

where n_{K} is the number of elements in partition ‘K’ and N is the total number of elements
CV_{K} =\sum_{K=1}^{K} MSE_{K}

CV_{K} =\frac{\sum_{K=1}^{K} MSE_{K}}{K}
Leave Out one Cross Validation (LOOCV) is a special case of K Fold Cross Validation where N-1 data points are used to train the model and 1 data point is used to test the model. There are N such paritions of N-1 & 1 that are possible. The mean error is measured The Cross Valifation Error for LOOCV is

CV_{N} = \frac{1}{n} *\frac{\sum_{1}^{n}(y-yhat)^{2}}{1-h_{i}}
where h_{i} is the diagonal hat matrix

see [Statistical Learning]

The above formula is also included in this blog post

It took me a day and a half to implement the K Fold Cross Validation formula. I think it is correct. In any case do let me know if you think it is off

5a. Leave out one cross validation (LOOCV) – R Code

R uses the package ‘boot’ for performing Cross Validation error computation

library(boot)
library(reshape2)
# Read data
df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))
# Select complete cases
df2 <- df1 %>% dplyr::select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]
set.seed(17)
cv.error=rep(0,10)
# For polynomials 1,2,3... 10 fit a LOOCV model
for (i in 1:10){
    glm.fit=glm(mpg~poly(horsepower,i),data=df3)
    cv.error[i]=cv.glm(df3,glm.fit)$delta[1]
    
}
cv.error
##  [1] 24.23151 19.24821 19.33498 19.42443 19.03321 18.97864 18.83305
##  [8] 18.96115 19.06863 19.49093
# Create and display a plot
folds <- seq(1,10)
df <- data.frame(folds,cvError=cv.error)
ggplot(df,aes(x=folds,y=cvError)) + geom_point() +geom_line(color="blue") +
    xlab("Degree of Polynomial") + ylab("Cross Validation Error") +
    ggtitle("Leave one out Cross Validation - Cross Validation Error vs Degree of Polynomial")

5b. Leave out one cross validation (LOOCV) – Python Code

In Python there is no available function to compute Cross Validation error and we have to compute the above formula. I have done this after several hours. I think it is now in reasonable shape. Do let me know if you think otherwise. For LOOCV I use the K Fold Cross Validation with K=N

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
from sklearn.cross_validation import train_test_split, KFold
from sklearn.preprocessing import PolynomialFeatures
from sklearn.metrics import mean_squared_error
# Read data
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
# Remove rows with NAs
autoDF3=autoDF2.dropna()
autoDF3.shape
X=autoDF3[['horsepower']]
y=autoDF3['mpg']

# For polynomial degree 1,2,3... 10
def computeCVError(X,y,folds):
    deg=[]
    mse=[]
    degree1=[1,2,3,4,5,6,7,8,9,10]
    
    nK=len(X)/float(folds)
    xval_err=0
    # For degree 'j'
    for j in degree1: 
        # Split as 'folds'
        kf = KFold(len(X),n_folds=folds)
        for train_index, test_index in kf:
            # Create the appropriate train and test partitions from the fold index
            X_train, X_test = X.iloc[train_index], X.iloc[test_index]
            y_train, y_test = y.iloc[train_index], y.iloc[test_index]  

            # For the polynomial degree 'j'
            poly = PolynomialFeatures(degree=j)        
            # Transform the X_train and X_test
            X_train_poly = poly.fit_transform(X_train)
            X_test_poly = poly.fit_transform(X_test)
            # Fit a model on the transformed data
            linreg = LinearRegression().fit(X_train_poly, y_train)
            # Compute yhat or ypred
            y_pred = linreg.predict(X_test_poly)   
            # Compute MSE * n_K/N
            test_mse = mean_squared_error(y_test, y_pred)*float(len(X_train))/float(len(X))     
            # Add the test_mse for this partition of the data
            mse.append(test_mse)
        # Compute the mean of all folds for degree 'j'   
        deg.append(np.mean(mse))
        
    return(deg)


df=pd.DataFrame()
print(len(X))
# Call the function once. For LOOCV K=N. hence len(X) is passed as number of folds
cvError=computeCVError(X,y,len(X))

# Create and plot LOOCV
df=pd.DataFrame(cvError)
fig3=df.plot()
fig3=plt.title("Leave one out Cross Validation - Cross Validation Error vs Degree of Polynomial")
fig3=plt.xlabel("Degree of Polynomial")
fig3=plt.ylabel("Cross validation Error")
fig3.figure.savefig('foo3.png', bbox_inches='tight')

 

6a K Fold Cross Validation – R code

Here K Fold Cross Validation is done for 4, 5 and 10 folds using the R package boot and the glm package

library(boot)
library(reshape2)
set.seed(17)
#Read data
df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))
df2 <- df1 %>% dplyr::select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]
a=matrix(rep(0,30),nrow=3,ncol=10)
set.seed(17)
# Set the folds as 4,5 and 10
folds<-c(4,5,10)
for(i in seq_along(folds)){
    cv.error.10=rep(0,10)
    for (j in 1:10){
        # Fit a generalized linear model
        glm.fit=glm(mpg~poly(horsepower,j),data=df3)
        # Compute K Fold Validation error
        a[i,j]=cv.glm(df3,glm.fit,K=folds[i])$delta[1]
        
    }
    
}

# Create and display the K Fold Cross Validation Error
b <- t(a)
df <- data.frame(b)
df1 <- cbind(seq(1,10),df)
names(df1) <- c("PolynomialDegree","4-fold","5-fold","10-fold")

df2 <- melt(df1,id="PolynomialDegree")
ggplot(df2) + geom_line(aes(x=PolynomialDegree, y=value, colour=variable),size=2) +
    xlab("Degree of Polynomial") + ylab("Cross Validation Error") +
    ggtitle("K Fold Cross Validation - Cross Validation Error vs Degree of Polynomial")

6b. K Fold Cross Validation – Python code

The implementation of K-Fold Cross Validation Error has to be implemented and I have done this below. There is a small discrepancy in the shapes of the curves with the R plot above. Not sure why!

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
from sklearn.cross_validation import train_test_split, KFold
from sklearn.preprocessing import PolynomialFeatures
from sklearn.metrics import mean_squared_error
# Read data
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
# Drop NA rows
autoDF3=autoDF2.dropna()
autoDF3.shape
#X=autoDF3[['cylinder','displacement','horsepower','weight']]
X=autoDF3[['horsepower']]
y=autoDF3['mpg']

# Create Cross Validation function
def computeCVError(X,y,folds):
    deg=[]
    mse=[]
    # For degree 1,2,3,..10
    degree1=[1,2,3,4,5,6,7,8,9,10]
    
    nK=len(X)/float(folds)
    xval_err=0
    for j in degree1: 
        # Split the data into 'folds'
        kf = KFold(len(X),n_folds=folds)
        for train_index, test_index in kf:
            # Partition the data acccording the fold indices generated
            X_train, X_test = X.iloc[train_index], X.iloc[test_index]
            y_train, y_test = y.iloc[train_index], y.iloc[test_index]  

            # Scale the X_train and X_test as per the polynomial degree 'j'
            poly = PolynomialFeatures(degree=j)             
            X_train_poly = poly.fit_transform(X_train)
            X_test_poly = poly.fit_transform(X_test)
            # Fit a polynomial regression
            linreg = LinearRegression().fit(X_train_poly, y_train)
            # Compute yhat or ypred
            y_pred = linreg.predict(X_test_poly)  
            # Compute MSE *(nK/N)
            test_mse = mean_squared_error(y_test, y_pred)*float(len(X_train))/float(len(X))  
            # Append to list for different folds
            mse.append(test_mse)
        # Compute the mean for poylnomial 'j' 
        deg.append(np.mean(mse))
        
    return(deg)

# Create and display a plot of K -Folds
df=pd.DataFrame()
for folds in [4,5,10]:
    cvError=computeCVError(X,y,folds)
    #print(cvError)
    df1=pd.DataFrame(cvError)
    df=pd.concat([df,df1],axis=1)
    #print(cvError)
    
df.columns=['4-fold','5-fold','10-fold']
df=df.reindex([1,2,3,4,5,6,7,8,9,10])
df
fig2=df.plot()
fig2=plt.title("K Fold Cross Validation - Cross Validation Error vs Degree of Polynomial")
fig2=plt.xlabel("Degree of Polynomial")
fig2=plt.ylabel("Cross validation Error")
fig2.figure.savefig('foo2.png', bbox_inches='tight')

output

This concludes this 2nd part of this series. I will look into model tuning and model selection in R and Python in the coming parts. Comments, suggestions and corrections are welcome!
To be continued….
Watch this space!

Also see

  1. Design Principles of Scalable, Distributed Systems
  2. Re-introducing cricketr! : An R package to analyze performances of cricketers
  3. Spicing up a IBM Bluemix cloud app with MongoDB and NodeExpress
  4. Using Linear Programming (LP) for optimizing bowling change or batting lineup in T20 cricket
  5. Simulating an Edge Shape in Android

To see all posts see Index of posts

Advertisements

Practical Machine Learning with R and Python – Part 1


Introduction

This is the 1st part of a series of posts I intend to write on some common Machine Learning Algorithms in R and Python. In this first part I cover the following Machine Learning Algorithms

  • Univariate Regression
  • Multivariate Regression
  • Polynomial Regression
  • K Nearest Neighbors Regression

The code includes the implementation in both R and Python. This series of posts are based on the following 2 MOOC courses I did at Stanford Online and at Coursera

  1. Statistical Learning, Prof Trevor Hastie & Prof Robert Tibesherani, Online Stanford
  2. Applied Machine Learning in Python Prof Kevyn-Collin Thomson, University Of Michigan, Coursera

I have used the data sets from UCI Machine Learning repository(Communities and Crime and Auto MPG). I also use the Boston data set from MASS package

While coding in R and Python I found that there were some aspects that were more convenient in one language and some in the other. For example, plotting the fit in R is straightforward in R, while computing the R squared, splitting as Train & Test sets etc. are already available in Python. In any case, these minor inconveniences can be easily be implemented in either language.

R squared computation in R is computed as follows
RSS=\sum (y-yhat)^{2}
TSS= \sum(y-mean(y))^{2}
Rsquared- 1-\frac{RSS}{TSS}

Note: You can download this R Markdown file and the associated data sets from Github at MachineLearning-RandPython
Note 1: This post was created as an R Markdown file in RStudio which has a cool feature of including R and Python snippets. The plot of matplotlib needs a workaround but otherwise this is a real cool feature of RStudio!

1.1a Univariate Regression – R code

Here a simple linear regression line is fitted between a single input feature and the target variable

# Source in the R function library
source("RFunctions.R")
# Read the Boston data file
df=read.csv("Boston.csv",stringsAsFactors = FALSE) # Data from MASS - Statistical Learning

# Split the data into training and test sets (75:25)
train_idx <- trainTestSplit(df,trainPercent=75,seed=5)
train <- df[train_idx, ]
test <- df[-train_idx, ]

# Fit a linear regression line between 'Median value of owner occupied homes' vs 'lower status of 
# population'
fit=lm(medv~lstat,data=df)
# Display details of fir
summary(fit)
## 
## Call:
## lm(formula = medv ~ lstat, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -15.168  -3.990  -1.318   2.034  24.500 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 34.55384    0.56263   61.41   <2e-16 ***
## lstat       -0.95005    0.03873  -24.53   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 6.216 on 504 degrees of freedom
## Multiple R-squared:  0.5441, Adjusted R-squared:  0.5432 
## F-statistic: 601.6 on 1 and 504 DF,  p-value: < 2.2e-16
# Display the confidence intervals
confint(fit)
##                 2.5 %     97.5 %
## (Intercept) 33.448457 35.6592247
## lstat       -1.026148 -0.8739505
plot(df$lstat,df$medv, xlab="Lower status (%)",ylab="Median value of owned homes ($1000)", main="Median value of homes ($1000) vs Lowe status (%)")
abline(fit)
abline(fit,lwd=3)
abline(fit,lwd=3,col="red")

rsquared=Rsquared(fit,test,test$medv)
sprintf("R-squared for uni-variate regression (Boston.csv)  is : %f", rsquared)
## [1] "R-squared for uni-variate regression (Boston.csv)  is : 0.556964"

1.1b Univariate Regression – Python code

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
#os.chdir("C:\\software\\machine-learning\\RandPython")

# Read the CSV file
df = pd.read_csv("Boston.csv",encoding = "ISO-8859-1")
# Select the feature variable
X=df['lstat']

# Select the target 
y=df['medv']

# Split into train and test sets (75:25)
X_train, X_test, y_train, y_test = train_test_split(X, y,random_state = 0)
X_train=X_train.values.reshape(-1,1)
X_test=X_test.values.reshape(-1,1)

# Fit a linear model
linreg = LinearRegression().fit(X_train, y_train)

# Print the training and test R squared score
print('R-squared score (training): {:.3f}'.format(linreg.score(X_train, y_train)))
print('R-squared score (test): {:.3f}'.format(linreg.score(X_test, y_test)))
     
# Plot the linear regression line
fig=plt.scatter(X_train,y_train)

# Create a range of points. Compute yhat=coeff1*x + intercept and plot
x=np.linspace(0,40,20)
fig1=plt.plot(x, linreg.coef_ * x + linreg.intercept_, color='red')
fig1=plt.title("Median value of homes ($1000) vs Lowe status (%)")
fig1=plt.xlabel("Lower status (%)")
fig1=plt.ylabel("Median value of owned homes ($1000)")
fig.figure.savefig('foo.png', bbox_inches='tight')
fig1.figure.savefig('foo1.png', bbox_inches='tight')
print "Finished"
## R-squared score (training): 0.571
## R-squared score (test): 0.458
## Finished

1.2a Multivariate Regression – R code

# Read crimes data
crimesDF <- read.csv("crimes.csv",stringsAsFactors = FALSE)

# Remove the 1st 7 columns which do not impact output
crimesDF1 <- crimesDF[,7:length(crimesDF)]

# Convert all to numeric
crimesDF2 <- sapply(crimesDF1,as.numeric)

# Check for NAs
a <- is.na(crimesDF2)
# Set to 0 as an imputation
crimesDF2[a] <-0
#Create as a dataframe
crimesDF2 <- as.data.frame(crimesDF2)
#Create a train/test split
train_idx <- trainTestSplit(crimesDF2,trainPercent=75,seed=5)
train <- crimesDF2[train_idx, ]
test <- crimesDF2[-train_idx, ]

# Fit a multivariate regression model between crimesPerPop and all other features
fit <- lm(ViolentCrimesPerPop~.,data=train)

# Compute and print R Squared
rsquared=Rsquared(fit,test,test$ViolentCrimesPerPop)
sprintf("R-squared for multi-variate regression (crimes.csv)  is : %f", rsquared)
## [1] "R-squared for multi-variate regression (crimes.csv)  is : 0.653940"

1.2b Multivariate Regression – Python code

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
# Read the data
crimesDF =pd.read_csv("crimes.csv",encoding="ISO-8859-1")
#Remove the 1st 7 columns
crimesDF1=crimesDF.iloc[:,7:crimesDF.shape[1]]
# Convert to numeric
crimesDF2 = crimesDF1.apply(pd.to_numeric, errors='coerce')
# Impute NA to 0s
crimesDF2.fillna(0, inplace=True)

# Select the X (feature vatiables - all)
X=crimesDF2.iloc[:,0:120]

# Set the target
y=crimesDF2.iloc[:,121]

X_train, X_test, y_train, y_test = train_test_split(X, y,random_state = 0)
# Fit a multivariate regression model
linreg = LinearRegression().fit(X_train, y_train)

# compute and print the R Square
print('R-squared score (training): {:.3f}'.format(linreg.score(X_train, y_train)))
print('R-squared score (test): {:.3f}'.format(linreg.score(X_test, y_test)))
## R-squared score (training): 0.699
## R-squared score (test): 0.677

1.3a Polynomial Regression – R

For Polynomial regression , polynomials of degree 1,2 & 3 are used and R squared is computed. It can be seen that the quadaratic model provides the best R squared score and hence the best fit

 # Polynomial degree 1
df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))

# Select key columns
df2 <- df1 %>% select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]

# Split as train and test sets
train_idx <- trainTestSplit(df3,trainPercent=75,seed=5)
train <- df3[train_idx, ]
test <- df3[-train_idx, ]

# Fit a model of degree 1
fit <- lm(mpg~. ,data=train)
rsquared1 <-Rsquared(fit,test,test$mpg)
sprintf("R-squared for Polynomial regression of degree 1 (auto_mpg.csv)  is : %f", rsquared1)
## [1] "R-squared for Polynomial regression of degree 1 (auto_mpg.csv)  is : 0.763607"
# Polynomial degree 2 - Quadratic
x = as.matrix(df3[1:6])
# Make a  polynomial  of degree 2 for feature variables before split
df4=as.data.frame(poly(x,2,raw=TRUE))
df5 <- cbind(df4,df3[7])

# Split into train and test set
train_idx <- trainTestSplit(df5,trainPercent=75,seed=5)
train <- df5[train_idx, ]
test <- df5[-train_idx, ]

# Fit the quadratic model
fit <- lm(mpg~. ,data=train)
# Compute R squared
rsquared2=Rsquared(fit,test,test$mpg)
sprintf("R-squared for Polynomial regression of degree 2 (auto_mpg.csv)  is : %f", rsquared2)
## [1] "R-squared for Polynomial regression of degree 2 (auto_mpg.csv)  is : 0.831372"
#Polynomial degree 3
x = as.matrix(df3[1:6])
# Make polynomial of degree 4  of feature variables before split
df4=as.data.frame(poly(x,3,raw=TRUE))
df5 <- cbind(df4,df3[7])
train_idx <- trainTestSplit(df5,trainPercent=75,seed=5)

train <- df5[train_idx, ]
test <- df5[-train_idx, ]
# Fit a model of degree 3
fit <- lm(mpg~. ,data=train)
# Compute R squared
rsquared3=Rsquared(fit,test,test$mpg)
sprintf("R-squared for Polynomial regression of degree 2 (auto_mpg.csv)  is : %f", rsquared3)
## [1] "R-squared for Polynomial regression of degree 2 (auto_mpg.csv)  is : 0.773225"
df=data.frame(degree=c(1,2,3),Rsquared=c(rsquared1,rsquared2,rsquared3))
# Make a plot of Rsquared and degree
ggplot(df,aes(x=degree,y=Rsquared)) +geom_point() + geom_line(color="blue") +
    ggtitle("Polynomial regression - R squared vs Degree of polynomial") +
    xlab("Degree") + ylab("R squared")

1.3a Polynomial Regression – Python

For Polynomial regression , polynomials of degree 1,2 & 3 are used and R squared is computed. It can be seen that the quadaratic model provides the best R squared score and hence the best fit

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
# Select key columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
# Convert columns to numeric
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
# Drop NAs
autoDF3=autoDF2.dropna()
autoDF3.shape
X=autoDF3[['cylinder','displacement','horsepower','weight','acceleration','year']]
y=autoDF3['mpg']

# Polynomial degree 1
X_train, X_test, y_train, y_test = train_test_split(X, y,random_state = 0)
linreg = LinearRegression().fit(X_train, y_train)
print('R-squared score - Polynomial degree 1 (training): {:.3f}'.format(linreg.score(X_train, y_train)))
# Compute R squared     
rsquared1 =linreg.score(X_test, y_test)
print('R-squared score - Polynomial degree 1 (test): {:.3f}'.format(linreg.score(X_test, y_test)))

# Polynomial degree 2
poly = PolynomialFeatures(degree=2)
X_poly = poly.fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X_poly, y,random_state = 0)
linreg = LinearRegression().fit(X_train, y_train)

# Compute R squared
print('R-squared score - Polynomial degree 2 (training): {:.3f}'.format(linreg.score(X_train, y_train)))
rsquared2 =linreg.score(X_test, y_test)
print('R-squared score - Polynomial degree 2 (test): {:.3f}\n'.format(linreg.score(X_test, y_test)))

#Polynomial degree 3

poly = PolynomialFeatures(degree=3)
X_poly = poly.fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X_poly, y,random_state = 0)
linreg = LinearRegression().fit(X_train, y_train)
print('(R-squared score -Polynomial degree 3  (training): {:.3f}'
     .format(linreg.score(X_train, y_train)))
# Compute R squared     
rsquared3 =linreg.score(X_test, y_test)
print('R-squared score Polynomial degree 3 (test): {:.3f}\n'.format(linreg.score(X_test, y_test)))
degree=[1,2,3]
rsquared =[rsquared1,rsquared2,rsquared3]
fig2=plt.plot(degree,rsquared)
fig2=plt.title("Polynomial regression - R squared vs Degree of polynomial")
fig2=plt.xlabel("Degree")
fig2=plt.ylabel("R squared")
fig2.figure.savefig('foo2.png', bbox_inches='tight')
print "Finished plotting and saving"
## R-squared score - Polynomial degree 1 (training): 0.811
## R-squared score - Polynomial degree 1 (test): 0.799
## R-squared score - Polynomial degree 2 (training): 0.861
## R-squared score - Polynomial degree 2 (test): 0.847
## 
## (R-squared score -Polynomial degree 3  (training): 0.933
## R-squared score Polynomial degree 3 (test): 0.710
## 
## Finished plotting and saving

1.4 K Nearest Neighbors

The code below implements KNN Regression both for R and Python. This is done for different neighbors. The R squared is computed in each case. This is repeated after performing feature scaling. It can be seen the model fit is much better after feature scaling. Normalization refers to

X_{normalized} = \frac{X-min(X)}{max(X-min(X))}

Another technique that is used is Standardization which is

X_{standardized} = \frac{X-mean(X)}{sd(X)}

1.4a K Nearest Neighbors Regression – R( Unnormalized)

The R code below does not use feature scaling

# KNN regression requires the FNN package
df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))
df2 <- df1 %>% select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]

# Split train and test
train_idx <- trainTestSplit(df3,trainPercent=75,seed=5)
train <- df3[train_idx, ]
test <- df3[-train_idx, ]
#  Select the feature variables
train.X=train[,1:6]
# Set the target for training
train.Y=train[,7]
# Do the same for test set
test.X=test[,1:6]
test.Y=test[,7]

rsquared <- NULL
# Create a list of neighbors
neighbors <-c(1,2,4,8,10,14)
for(i in seq_along(neighbors)){
    # Perform a KNN regression fit
    knn=knn.reg(train.X,test.X,train.Y,k=neighbors[i])
    # Compute R sqaured
    rsquared[i]=knnRSquared(knn$pred,test.Y)
}

# Make a dataframe for plotting
df <- data.frame(neighbors,Rsquared=rsquared)
# Plot the number of neighors vs the R squared
ggplot(df,aes(x=neighbors,y=Rsquared)) + geom_point() +geom_line(color="blue") +
    xlab("Number of neighbors") + ylab("R squared") +
    ggtitle("KNN regression - R squared vs Number of Neighors (Unnormalized)")

1.4b K Nearest Neighbors Regression – Python( Unnormalized)

The Python code below does not use feature scaling

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
from sklearn.neighbors import KNeighborsRegressor
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
autoDF3=autoDF2.dropna()
autoDF3.shape
X=autoDF3[['cylinder','displacement','horsepower','weight','acceleration','year']]
y=autoDF3['mpg']

# Perform a train/test split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state = 0)
# Create a list of neighbors
rsquared=[]
neighbors=[1,2,4,8,10,14]
for i in neighbors:
        # Fit a KNN model
        knnreg = KNeighborsRegressor(n_neighbors = i).fit(X_train, y_train)
        # Compute R squared
        rsquared.append(knnreg.score(X_test, y_test))
        print('R-squared test score: {:.3f}'
        .format(knnreg.score(X_test, y_test)))
# Plot the number of neighors vs the R squared        
fig3=plt.plot(neighbors,rsquared)
fig3=plt.title("KNN regression - R squared vs Number of neighbors(Unnormalized)")
fig3=plt.xlabel("Neighbors")
fig3=plt.ylabel("R squared")
fig3.figure.savefig('foo3.png', bbox_inches='tight')
print "Finished plotting and saving"
## R-squared test score: 0.527
## R-squared test score: 0.678
## R-squared test score: 0.707
## R-squared test score: 0.684
## R-squared test score: 0.683
## R-squared test score: 0.670
## Finished plotting and saving

1.4c K Nearest Neighbors Regression – R( Normalized)

It can be seen that R squared improves when the features are normalized.

df=read.csv("auto_mpg.csv",stringsAsFactors = FALSE) # Data from UCI
df1 <- as.data.frame(sapply(df,as.numeric))
df2 <- df1 %>% select(cylinder,displacement, horsepower,weight, acceleration, year,mpg)
df3 <- df2[complete.cases(df2),]

# Perform MinMaxScaling of feature variables 
train.X.scaled=MinMaxScaler(train.X)
test.X.scaled=MinMaxScaler(test.X)

# Create a list of neighbors
rsquared <- NULL
neighbors <-c(1,2,4,6,8,10,12,15,20,25,30)
for(i in seq_along(neighbors)){
    # Fit a KNN model
    knn=knn.reg(train.X.scaled,test.X.scaled,train.Y,k=i)
    # Compute R ssquared
    rsquared[i]=knnRSquared(knn$pred,test.Y)
    
}

df <- data.frame(neighbors,Rsquared=rsquared)
# Plot the number of neighors vs the R squared 
ggplot(df,aes(x=neighbors,y=Rsquared)) + geom_point() +geom_line(color="blue") +
    xlab("Number of neighbors") + ylab("R squared") +
    ggtitle("KNN regression - R squared vs Number of Neighors(Normalized)")

1.4d K Nearest Neighbors Regression – Python( Normalized)

R squared improves when the features are normalized with MinMaxScaling

import numpy as np
import pandas as pd
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
from sklearn.neighbors import KNeighborsRegressor
from sklearn.preprocessing import MinMaxScaler
autoDF =pd.read_csv("auto_mpg.csv",encoding="ISO-8859-1")
autoDF.shape
autoDF.columns
autoDF1=autoDF[['mpg','cylinder','displacement','horsepower','weight','acceleration','year']]
autoDF2 = autoDF1.apply(pd.to_numeric, errors='coerce')
autoDF3=autoDF2.dropna()
autoDF3.shape
X=autoDF3[['cylinder','displacement','horsepower','weight','acceleration','year']]
y=autoDF3['mpg']

# Perform a train/ test  split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state = 0)
# Use MinMaxScaling
scaler = MinMaxScaler()
X_train_scaled = scaler.fit_transform(X_train)
# Apply scaling on test set
X_test_scaled = scaler.transform(X_test)

# Create a list of neighbors
rsquared=[]
neighbors=[1,2,4,6,8,10,12,15,20,25,30]
for i in neighbors:
    # Fit a KNN model
    knnreg = KNeighborsRegressor(n_neighbors = i).fit(X_train_scaled, y_train)
    # Compute R squared
    rsquared.append(knnreg.score(X_test_scaled, y_test))
    print('R-squared test score: {:.3f}'
        .format(knnreg.score(X_test_scaled, y_test)))

# Plot the number of neighors vs the R squared 
fig4=plt.plot(neighbors,rsquared)
fig4=plt.title("KNN regression - R squared vs Number of neighbors(Normalized)")
fig4=plt.xlabel("Neighbors")
fig4=plt.ylabel("R squared")
fig4.figure.savefig('foo4.png', bbox_inches='tight')
print "Finished plotting and saving"
## R-squared test score: 0.703
## R-squared test score: 0.810
## R-squared test score: 0.830
## R-squared test score: 0.838
## R-squared test score: 0.834
## R-squared test score: 0.828
## R-squared test score: 0.827
## R-squared test score: 0.826
## R-squared test score: 0.816
## R-squared test score: 0.815
## R-squared test score: 0.809
## Finished plotting and saving

Conclusion

In this initial post I cover the regression models when the output is continous. I intend to touch upon other Machine Learning algorithms.
Comments, suggestions and corrections are welcome.

Watch this this space!

To be continued….

You may like
1. Using Linear Programming (LP) for optimizing bowling change or batting lineup in T20 cricket
2. Neural Networks: The mechanics of backpropagation
3. More book, more cricket! 2nd edition of my books now on Amazon
4. Spicing up a IBM Bluemix cloud app with MongoDB and NodeExpress
5. Introducing cricket package yorkr:Part 4-In the block hole!

To see all posts see Index of posts

Using Linear Programming (LP) for optimizing bowling change or batting lineup in T20 cricket


In my recent post, My travels through the realms of Data Science, Machine Learning, Deep Learning and (AI), I had recounted my journey in the domains of of Data Science, Machine Learning (ML), and more recently Deep Learning (DL) all of which are useful while analyzing data. Of late, I have come to the realization that there are many facets to data. And to glean insights from data, Data Science, ML and DL alone are not sufficient and one needs to also have a good handle on linear programming and optimization. My colleague at IBM Research also concurred with this view and told me he had arrived at this conclusion several years ago.

While ML & DL are very useful and interesting to make inferences and predictions of outputs from input variables, optimization computes the choice of input which results in maximizing or minimizing the output. So I made a small course correction and started on a course from India’s own NPTEL Introduction to Linear Programming by Prof G. Srinivasan of IIT Madras (highly recommended!). The lectures are delivered with remarkable clarity by the Prof and I am just about halfway through the course (each lecture is of 50-55 min duration), when I decided that I needed to try to formulate and solve some real world Linear Programming problem.

As usual, I turned towards cricket for some appropriate situations, and sure enough it was there in the open. For this LP formulation I take International T20 and IPL, though International ODI will also work equally well.  You can download the associated code and data for this from Github at LP-cricket-analysis

In T20 matches the captain has to make choice of how to rotate bowlers with the aim of restricting the batting side. Conversely, the batsmen need to take advantage of the bowling strength to maximize the runs scored.

Note:
a) A simple and obvious strategy would be
– If the ith bowler’s economy rate is less than the economy rate of the jth bowler i.e.
er_{i} < er_{j} then have bowler ‘i’ to bowl more overs as his/her economy rate is better

b)A better strategy would be to consider the economy rate of each bowler against each batsman. How often  have we witnessed bowlers with a great bowling average get thrashed time and again by the same batsman, or a bowler who is generally very poor being very effective against a particular batsman. i.e. er_{ij} < er_{ik} where the jth bowler is more effective than the kth bowler against the ith batsman. This now becomes a linear optimization problem as we can have several combinations of number of overs X economy rate for different bowlers and we will have to solve this algorithmically to determine the lowest score for bowling performance or highest score for batting order.

This post uses the latter approach to optimize bowling change and batting lineup.

Let is take a hypothetical situation
Assume there are 3 bowlers – bwlr_{1},bwlr_{2},bwlr_{3}
and there are 4 batsmen – bman_{1},bman_{2},bman_{3},bman_{4}

Let the economy rate er_{ij} be the Economy Rate of the jth bowler to the ith batsman. Also if remaining overs for the bowlers are o_{1},o_{2},o_{3}
and the total number of overs left to be bowled are
o_{1}+o_{2}+o_{3} = N then the question is

a) Given the economy rate of each bowler per batsman, how many overs should each bowler bowl, so that the total runs scored by all the batsmen are minimum?

b) Alternatively, if the know the individual strike rate of a batsman against the individual bowlers, how many overs should each batsman face with a bowler so that the total runs scored is maximized?

1. LP Formulation for bowling order

Let the economy rate er_{ij} be the Economy Rate of the jth bowler to the ith batsman.
Objective function : Minimize –
er_{11}*o_{11} + er_{12}*o_{12} +..+er_{1n}*o_{1n}+ er_{21}*o_{21} + er_{22}*o_{22}+.. + er_{22}*o_{2n}+ er_{m1}*o_{m1}+..+ er_{mn}*o_{mn}
i.e.
\sum_{i=1}^{i=m}\sum_{j=1}^{i=n}er_{ij}*o_{ij}
Constraints
Where o_{j} is the number of overs remaining for the jth bowler against  ‘k’ batsmen
o_{j1} + o_{j2} + .. o_{jk} < o_{j}
and if the total number of overs remaining to be bowled is N then
o_{1} + o_{2} +...+ o_{k} = N or
\sum_{j=1}^{j=k} o_{j} =N
The overs that any bowler can bowl is o_{j} >=0

2. LP Formulation for batting lineup

Let the strike rate sr_{ij}  be the Strike Rate of the ith batsman to the jth bowler
Objective function : Maximize –
sr_{11}*o_{11} + sr_{12}*o_{12} +..+ sr_{1n}*o_{1n}+ sr_{21}*o_{21} + sr_{22}*o_{22}+.. sr_{2n}*o_{2n}+ sr_{m1}*o_{m1}+..+ sr_{mn}*o_{mn}
i.e.
\sum_{i=1}^{i=4}\sum_{j=1}^{i=3}sr_{ij}*o_{ij}
Constraints
Where o_{j} is the number of overs remaining for the jth bowler against  ‘k’ batsmen
o_{j1} + o_{j2} + .. o_{jk} < o_{j}
and the total number of overs remaining to be bowled is N then
o_{1} + o_{2} +...+ o_{k} = N or
\sum_{j=1}^{j=k} o_{j} =N
The overs that any bowler can bowl is
o_{j} >=0

lpSolveAPI– For this maximization and minimization problem I used lpSolveAPI.

Below I take 2 simple examples (example1 & 2)  to ensure that my LP formulation and solution is correct before applying it on real T20 cricket data (Intl. T20 and IPL)

3. LP formulation (Example 1)

Initially I created a test example to ensure that I get the LP formulation and solution correct. Here the er1=4 and er2=3 and o1 & o2 are the overs bowled by bowlers 1 & 2. Also o1+o2=4 In this example as below

o1 o2 Obj Fun(=4o1+3o2)
1    3      13
2    2      14
3    1      15

library(lpSolveAPI)
library(dplyr)
library(knitr)
lprec <- make.lp(0, 2)
a <-lp.control(lprec, sense="min")
set.objfn(lprec, c(4, 3))  # Economy Rate of 4 and 3 for er1 and er2
add.constraint(lprec, c(1, 1), "=",4)  # o1 + o2 =4
add.constraint(lprec, c(1, 0), ">",1)  # o1 > 1
add.constraint(lprec, c(0, 1), ">",1)  # o2 > 1
lprec
## Model name: 
##             C1    C2       
## Minimize     4     3       
## R1           1     1   =  4
## R2           1     0  >=  1
## R3           0     1  >=  1
## Kind       Std   Std       
## Type      Real  Real       
## Upper      Inf   Inf       
## Lower        0     0
b <-solve(lprec)
get.objective(lprec) # 13
## [1] 13
get.variables(lprec) # 1    3 
## [1] 1 3

Note 1: In the above example 13 runs is the minimum that can be scored and this requires

LP solution:
Minimum runs=13

  • o1=1
  • o2=3

Note 2:The numbers in the columns represent the number of overs that need to be bowled by a bowler to the corresponding batsman.

4. LP formulation (Example 2)

In this formulation there are 2 bowlers and 2 batsmen o11,o12 are the oves bowled by bowler 1 to batsmen 1 & 2 and o21, o22 are the overs bowled by bowler 2 to batsmen 1 & 2 er11=4, er12=2,er21=2,er22=5 o11+o12+o21+o22=5

The solution for this manually computed is o11, o12, o21, o22 Runs
where B11, B12 are the overs bowler 1 bowls to batsman 1 and B21 and B22 are overs bowler 2 bowls to batsman 2

o11     o12    o21    o22      Runs=(4*o11+2*o12+2*o21+5*o22)
1            1             1            2           18
1           2              1             1           15
2           1              1            1            17
1           1               2            1            15

lprec <- make.lp(0, 4)
a <-lp.control(lprec, sense="min")
set.objfn(lprec, c(4, 2,2,5))
add.constraint(lprec, c(1, 1,0,0), "<=",8)
add.constraint(lprec, c(0, 0,1,1), "<=",7)
add.constraint(lprec, c(1, 1,1,1), "=",5)
add.constraint(lprec, c(1, 0,0,0), ">",1)
add.constraint(lprec, c(0, 1,0,0), ">",1)
add.constraint(lprec, c(0, 0,1,0), ">",1)
add.constraint(lprec, c(0, 0,0,1), ">",1)
lprec
## Model name: 
##             C1    C2    C3    C4       
## Minimize     4     2     2     5       
## R1           1     1     0     0  <=  8
## R2           0     0     1     1  <=  7
## R3           1     1     1     1   =  5
## R4           1     0     0     0  >=  1
## R5           0     1     0     0  >=  1
## R6           0     0     1     0  >=  1
## R7           0     0     0     1  >=  1
## Kind       Std   Std   Std   Std       
## Type      Real  Real  Real  Real       
## Upper      Inf   Inf   Inf   Inf       
## Lower        0     0     0     0
b<-solve(lprec)
get.objective(lprec) 
## [1] 15
get.variables(lprec) 
## [1] 1 2 1 1

Note: In the above example 15 runs is the minimum that can be scored and this requires

LP Solution:
Minimum runs=15

  • o11=1
  • o12=2
  • o21=1
  • o22=1

It is possible to keep the minimum to other values and solves also.

5. LP formulation for International T20 India vs Australia (Batting lineup)

To analyze batting and bowling lineups in the cricket world I needed to get the ball-by-ball details of runs scored by each batsman against each of the bowlers. Fortunately I had already created this with my R package yorkr. yorkr processes yaml data from Cricsheet. So I copied the data of all matches between Australia and India in International T20s. You can download my processed data for International T20 at Inswinger

load("Australia-India-allMatches.RData")
dim(matches)
## [1] 3541   25

The following functions compute the ‘Strike Rate’ of a batsman as

SR=1/oversRunsScored

Also the Economy Rate is computed as

ER=1/oversRunsConceded

Incidentally the SR=ER

# Compute the Strike Rate of the batsman
computeSR <- function(batsman1,bowler1){
    a <- matches %>% filter(batsman==batsman1 & bowler==bowler1) 
    a1 <- a %>% summarize(totalRuns=sum(runs),count=n()) %>% mutate(SR=(totalRuns/count)*6)
    a1
}

# Compute the Economy Rate of the batsman
computeER <- function(batsman1,bowler1){
    a <- matches %>% filter(batsman==batsman1 & bowler==bowler1) 
    a1 <- a %>% summarize(totalRuns=sum(runs),count=n()) %>% mutate(ER=(totalRuns/count)*6)
    a1
}

Here I compute the Strike Rate of Virat Kohli, Yuvraj Singh and MS Dhoni against Shane Watson, Brett Lee and MA Starc

 # Kohli
kohliWatson<- computeSR("V Kohli","SR Watson")
kohliWatson
##   totalRuns count       SR
## 1        45    37 7.297297
kohliLee <- computeSR("V Kohli","B Lee")
kohliLee
##   totalRuns count       SR
## 1        10     7 8.571429
kohliStarc <- computeSR("V Kohli","MA Starc")
kohliStarc
##   totalRuns count       SR
## 1        11     9 7.333333
# Yuvraj
yuvrajWatson<- computeSR("Yuvraj Singh","SR Watson")
yuvrajWatson
##   totalRuns count       SR
## 1        24    22 6.545455
yuvrajLee <- computeSR("Yuvraj Singh","B Lee")
yuvrajLee
##   totalRuns count       SR
## 1        12     7 10.28571
yuvrajStarc <- computeSR("Yuvraj Singh","MA Starc")
yuvrajStarc
##   totalRuns count SR
## 1        12     8  9
# MS Dhoni
dhoniWatson<- computeSR("MS Dhoni","SR Watson")
dhoniWatson
##   totalRuns count       SR
## 1        33    28 7.071429
dhoniLee <- computeSR("MS Dhoni","B Lee")
dhoniLee
##   totalRuns count  SR
## 1        26    20 7.8
dhoniStarc <- computeSR("MS Dhoni","MA Starc")
dhoniStarc
##   totalRuns count   SR
## 1        11     8 8.25

When we consider the batting lineup, the problem is one of maximization. In the LP formulation below V Kohli has a SR of 7.29, 8.57, 7.33 against Watson, Lee & Starc
Yuvraj has a SR of 6.5, 10.28, 9 against Watson, Lee & Starc
and Dhoni has a SR of 7.07, 7.8,  8.25 against Watson, Lee and Starc

The constraints are Watson, Lee and Starc have 3, 4 & 3 overs remaining respectively. The total number of overs remaining to be bowled is 9.The other constraints could be that a bowler bowls at least 1 over etc.

Formulating and solving

# 3 batsman x 3 bowlers
lprec <- make.lp(0, 9)
# Maximization
a<-lp.control(lprec, sense="max")

# Set the objective function
set.objfn(lprec, c(kohliWatson$SR, kohliLee$SR,kohliStarc$SR,
                   yuvrajWatson$SR,yuvrajLee$SR,yuvrajStarc$SR,
                   dhoniWatson$SR,dhoniLee$SR,dhoniStarc$SR))

#Assume the  bowlers have 3,4,3 overs left respectively
add.constraint(lprec, c(1, 1,1,0,0,0, 0,0,0), "<=",3)
add.constraint(lprec, c(0,0,0,1,1,1,0,0,0), "<=",4)
add.constraint(lprec, c(0,0,0,0,0,0,1,1,1), "<=",3)
#o11+o12+o13+o21+o22+o23+o31+o32+o33=8 (overs remaining)
add.constraint(lprec, c(1,1,1,1,1,1,1,1,1), "=",9) 


add.constraint(lprec, c(1,0,0,0,0,0,0,0,0), ">=",1) #o11 >=1
add.constraint(lprec, c(0,1,0,0,0,0,0,0,0), ">=",0) #o12 >=0
add.constraint(lprec, c(0,0,1,0,0,0,0,0,0), ">=",0) #o13 >=0
add.constraint(lprec, c(0,0,0,1,0,0,0,0,0), ">=",1) #o21 >=1
add.constraint(lprec, c(0,0,0,0,1,0,0,0,0), ">=",1) #o22 >=1
add.constraint(lprec, c(0,0,0,0,0,1,0,0,0), ">=",0) #o23 >=0
add.constraint(lprec, c(0,0,0,0,0,0,1,0,0), ">=",1) #o31 >=1
add.constraint(lprec, c(0,0,0,0,0,0,0,1,0), ">=",0) #o32 >=0
add.constraint(lprec, c(0,0,0,0,0,0,0,0,1), ">=",0) #o33 >=0

lprec
## Model name: 
##   a linear program with 9 decision variables and 13 constraints
b <-solve(lprec)
get.objective(lprec) #  
## [1] 77.16418
get.variables(lprec) # 
## [1] 1 2 0 1 3 0 1 0 1

This shows that the maximum runs that can be scored for the current strike rate is 77.16   runs in 9 overs The breakup is as follows

This is also shown below

get.variables(lprec) # 
## [1] 1 2 0 1 3 0 1 0 1

This is also shown below

e <- as.data.frame(rbind(c(1,2,0,3),c(1,3,0,4),c(1,0,1,2)))
names(e) <- c("S Watson","B Lee","MA Starc","Overs")
rownames(e) <- c("Kohli","Yuvraj","Dhoni")
e

LP Solution:
Maximum runs that can be scored by India against Australia is:77.164 if the 9 overs to be faced by the batsman are as below

##        S Watson B Lee MA Starc Overs
## Kohli         1     2        0     3
## Yuvraj        1     3        0     4
## Dhoni         1     0        1     2
#Total overs=9

Note: This assumes that the batsmen perform at their current Strike Rate. Howvever anything can happen in a real game, but nevertheless this is a fairly reasonable estimate of the performance

Note 2:The numbers in the columns represent the number of overs that need to be bowled by a bowler to the corresponding batsman.

Note 3:You could try other combinations of overs for the above SR. For the above constraints 77.16 is the highest score for the given number of overs

6. LP formulation for International T20 India vs Australia (Bowling lineup)

For this I compute how the bowling should be rotated between R Ashwin, RA Jadeja and JJ Bumrah when taking into account their performance against batsmen like Shane Watson, AJ Finch and David Warner. For the bowling performance I take the Economy rate of the bowlers. The data is the same as above

computeSR <- function(batsman1,bowler1){
    a <- matches %>% filter(batsman==batsman1 & bowler==bowler1) 
    a1 <- a %>% summarize(totalRuns=sum(runs),count=n()) %>% mutate(SR=(totalRuns/count)*6)
    a1
}
# RA Jadeja
jadejaWatson<- computeER("SR Watson","RA Jadeja")
jadejaWatson
##   totalRuns count       ER
## 1        60    29 12.41379
jadejaFinch <- computeER("AJ Finch","RA Jadeja")
jadejaFinch
##   totalRuns count       ER
## 1        36    33 6.545455
jadejaWarner <- computeER("DA Warner","RA Jadeja")
jadejaWarner
##   totalRuns count       ER
## 1        23    11 12.54545
# Ashwin
ashwinWatson<- computeER("SR Watson","R Ashwin")
ashwinWatson
##   totalRuns count       ER
## 1        41    26 9.461538
ashwinFinch <- computeER("AJ Finch","R Ashwin")
ashwinFinch
##   totalRuns count   ER
## 1        63    36 10.5
ashwinWarner <- computeER("DA Warner","R Ashwin")
ashwinWarner
##   totalRuns count       ER
## 1        38    28 8.142857
# JJ Bunrah
bumrahWatson<- computeER("SR Watson","JJ Bumrah")
bumrahWatson
##   totalRuns count  ER
## 1        22    20 6.6
bumrahFinch <- computeER("AJ Finch","JJ Bumrah")
bumrahFinch
##   totalRuns count       ER
## 1        25    19 7.894737
bumrahWarner <- computeER("DA Warner","JJ Bumrah")
bumrahWarner
##   totalRuns count ER
## 1         2     4  3

As can be seen from above RA Jadeja has a ER of 12.4, 6.54, 12.54 against Watson, AJ Finch and Warner also Ashwin has a ER of 9.46, 10.5, 8.14 against Watson, Finch and Warner. Similarly Bumrah has an ER of 6.6,7.89, 3 against Watson, Finch and Warner
The constraints are Jadeja, Ashwin and Bumrah have 4, 3 & 4 overs remaining and the total overs remaining to be bowled is 10.

Formulating solving the bowling lineup is shown below

lprec <- make.lp(0, 9)
a <-lp.control(lprec, sense="min")

# Set the objective function
set.objfn(lprec, c(jadejaWatson$ER, jadejaFinch$ER,jadejaWarner$ER,
                   ashwinWatson$ER,ashwinFinch$ER,ashwinWarner$ER,
                   bumrahWatson$ER,bumrahFinch$ER,bumrahWarner$ER))

add.constraint(lprec, c(1, 1,1,0,0,0, 0,0,0), "<=",4) # Jadeja has 4 overs
add.constraint(lprec, c(0,0,0,1,1,1,0,0,0), "<=",3)   # Ashwin has 3 overs left
add.constraint(lprec, c(0,0,0,0,0,0,1,1,1), "<=",4)   # Bumrah has 4 overs left
add.constraint(lprec, c(1,1,1,1,1,1,1,1,1), "=",10) # Total overs = 10
add.constraint(lprec, c(1,0,0,0,0,0,0,0,0), ">=",1)
add.constraint(lprec, c(0,1,0,0,0,0,0,0,0), ">=",0)
add.constraint(lprec, c(0,0,1,0,0,0,0,0,0), ">=",1)
add.constraint(lprec, c(0,0,0,1,0,0,0,0,0), ">=",0)
add.constraint(lprec, c(0,0,0,0,1,0,0,0,0), ">=",1)
add.constraint(lprec, c(0,0,0,0,0,1,0,0,0), ">=",0)
add.constraint(lprec, c(0,0,0,0,0,0,1,0,0), ">=",0)
add.constraint(lprec, c(0,0,0,0,0,0,0,1,0), ">=",1)
add.constraint(lprec, c(0,0,0,0,0,0,0,0,1), ">=",0)

lprec
## Model name: 
##   a linear program with 9 decision variables and 13 constraints
b <-solve(lprec)
get.objective(lprec) #  
## [1] 73.58775
get.variables(lprec) # 
## [1] 1 2 1 0 1 1 0 1 3

The minimum runs that will be conceded by these 3 bowlers in 10 overs is 73.58 assuming the bowling is rotated as follows

e <- as.data.frame(rbind(c(1,0,0),c(2,1,1),c(1,1,3),c(4,2,4)))
names(e) <- c("RA Jadeja","R Ashwin","JJ Bumrah")
rownames(e) <- c("S Watson","AJ Finch","DA Warner","Overs")
e 

LP Solution:
Minimum runs that will be conceded by India against Australia is 73.58 in 10 overs if the overs bowled are as follows

##           RA Jadeja R Ashwin JJ Bumrah
## S Watson          1        0         0
## AJ Finch          2        1         1
## DA Warner         1        1         3
## Overs             4        2         4
#Total overs=10  

7. LP formulation for IPL (Mumbai Indians – Kolkata Knight Riders – Bowling lineup)

As in the case of International T20s I also have processed IPL data derived from my R package yorkr. yorkr. yorkr processes yaml data from Cricsheet. The processed data for all IPL matches can be downloaded from GooglyPlus

load("Mumbai Indians-Kolkata Knight Riders-allMatches.RData")
dim(matches)
## [1] 4237   25
# Compute the Economy Rate of the Mumbai Indian bowlers against Kolkata Knight Riders

# Gambhir
gambhirMalinga <- computeER("G Gambhir","SL Malinga")
gambhirHarbhajan <- computeER("G Gambhir","Harbhajan Singh")
gambhirPollard <- computeER("G Gambhir","KA Pollard")

#Yusuf Pathan
yusufMalinga <- computeER("YK Pathan","SL Malinga")
yusufHarbhajan <- computeER("YK Pathan","Harbhajan Singh")
yusufPollard <- computeER("YK Pathan","KA Pollard")

#JH Kallis
kallisMalinga <- computeER("JH Kallis","SL Malinga")
kallisHarbhajan <- computeER("JH Kallis","Harbhajan Singh")
kallisPollard <- computeER("JH Kallis","KA Pollard")

#RV Uthappa
uthappaMalinga <- computeER("RV Uthappa","SL Malinga")
uthappaHarbhajan <- computeER("RV Uthappa","Harbhajan Singh")
uthappaPollard <- computeER("RV Uthappa","KA Pollard")

Here

gambhirMalinga, yusufMalinga, kallisMalinga, uthappaMalinga is the ER of Malinga against Gambhir, Yusuf Pathan, Kallis and Uthappa
gambhirHarbhajan, yusufHarbhajan, kallisHarbhajan, uthappaHarbhajan is the ER of Harbhajan against Gambhir, Yusuf Pathan, Kallis and Uthappa
gambhirPollard, yusufPollard, kallisPollard, uthappaPollard is the ER of Kieron Pollard against Gambhir, Yusuf Pathan, Kallis and Uthappa

The constraints are Malinga, Harbhajan and Pollard have 4 overs each and remaining overs to be bowled is 10.

Formulating and solving this for the bowling lineup of Mumbai Indians against Kolkata Knight Riders

 library("lpSolveAPI")
 lprec <- make.lp(0, 12)
 a=lp.control(lprec, sense="min")
 
 set.objfn(lprec, c(gambhirMalinga$ER, yusufMalinga$ER,kallisMalinga$ER,uthappaMalinga$ER,
                    gambhirHarbhajan$ER,yusufHarbhajan$ER,kallisHarbhajan$ER,uthappaHarbhajan$ER,
                    gambhirPollard$ER,yusufPollard$ER,kallisPollard$ER,uthappaPollard$ER))
 
 add.constraint(lprec, c(1,1,1,1, 0,0,0,0, 0,0,0,0), "<=",4)
 add.constraint(lprec, c(0,0,0,0,1,1,1,1,0,0,0,0), "<=",4)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,1,1,1,1), "<=",4)
 add.constraint(lprec, c(1,1,1,1,1,1,1,1,1,1,1,1), "=",10)
 
 add.constraint(lprec, c(1,0,0,0,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,1,0,0,0,0,0,0,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,1,0,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,1,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,1,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,1,0,0,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,1,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,1,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,1,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,1,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,0,1,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,0,0,1), ">=",0)
 
 lprec
## Model name: 
##   a linear program with 12 decision variables and 16 constraints
 b=solve(lprec)
 get.objective(lprec) #  
## [1] 55.57887
 get.variables(lprec) # 
##  [1] 3 1 0 0 0 1 0 1 3 1 0 0
e <- as.data.frame(rbind(c(3,1,0,0,4),c(0, 1, 0,1,2),c(3, 1, 0,0,4)))
names(e) <- c("Gambhir","Yusuf","Kallis","Uthappa","Overs")
rownames(e) <- c("Malinga","Harbhajan","Pollard") 
e

LP Solution: Mumbai Indians can restrict Kolkata Knight Riders to 55.87 in 10 overs
if the overs are bowled as below

##           Gambhir Yusuf Kallis Uthappa Overs
## Malinga         3     1      0       0     4
## Harbhajan       0     1      0       1     2
## Pollard         3     1      0       0     4
#Total overs=10  

8. LP formulation for IPL (Mumbai Indians – Kolkata Knight Riders – Batting lineup)

As I mentioned it is possible to perform a maximation with the same formulation since computeSR<==>computeER

This just flips the problem around and computes the maximum runs that can be scored for the batsman’s Strike rate (this is same as the bowler’s Economy rate) i.e.

gambhirMalinga, yusufMalinga, kallisMalinga, uthappaMalinga is the SR of Gambhir, Yusuf Pathan, Kallis and Uthappa against Malinga
gambhirHarbhajan, yusufHarbhajan, kallisHarbhajan, uthappaHarbhajan is the SR of Gambhir, Yusuf Pathan, Kallis and Uthappa against Harbhajan
gambhirPollard, yusufPollard, kallisPollard, uthappaPollard is the SR of Gambhir, Yusuf Pathan, Kallis and Uthappa against Kieron Pollard.

The constraints are Malinga, Harbhajan and Pollard have 4 overs each and remaining overs to be bowled is 10.

 library("lpSolveAPI")
 lprec <- make.lp(0, 12)
 a=lp.control(lprec, sense="max")
 
 a <-set.objfn(lprec, c(gambhirMalinga$ER, yusufMalinga$ER,kallisMalinga$ER,uthappaMalinga$ER,
                    gambhirHarbhajan$ER,yusufHarbhajan$ER,kallisHarbhajan$ER,uthappaHarbhajan$ER,
                    gambhirPollard$ER,yusufPollard$ER,kallisPollard$ER,uthappaPollard$ER))
 
 
 add.constraint(lprec, c(1,1,1,1, 0,0,0,0, 0,0,0,0), "<=",4)
 add.constraint(lprec, c(0,0,0,0,1,1,1,1,0,0,0,0), "<=",4)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,1,1,1,1), "<=",4)
 add.constraint(lprec, c(1,1,1,1,1,1,1,1,1,1,1,1), "=",11)
 
 add.constraint(lprec, c(1,0,0,0,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,1,0,0,0,0,0,0,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,1,0,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,1,0,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,1,0,0,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,1,0,0,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,1,0,0,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,1,0,0,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,1,0,0,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,1,0,0), ">=",1)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,0,1,0), ">=",0)
 add.constraint(lprec, c(0,0,0,0,0,0,0,0,0,0,0,1), ">=",0)
 lprec
## Model name: 
##   a linear program with 12 decision variables and 16 constraints
 b=solve(lprec)
 get.objective(lprec) #  
## [1] 94.22649
 get.variables(lprec) # 
##  [1] 0 3 0 0 0 1 0 3 0 1 3 0
e <- as.data.frame(rbind(c(0,3,0,0,3),c(0, 1, 0,3,4),c(0, 1, 3,0,4)))
names(e) <- c("Gambhir","Yusuf","Kallis","Uthappa","Overs")
rownames(e) <- c("Malinga","Harbhajan","Pollard") 
e

LP Solution: Kolkata Knight Riders can score a maximum of 94.22 in 11 overs against Mumbai Indians
if the the number of overs KKR face is as below

##           Gambhir Yusuf Kallis Uthappa Overs
## Malinga         0     3      0       0     3
## Harbhajan       0     1      0       3     4
## Pollard         0     1      3       0     4
#Total overs=11  

Conclusion: It is possible to thus determine the optimum no of overs to give to a specific bowler based on his/her Economy Rate with a particular batsman. Similarly one can determine the maximum runs that can be scored by a batsmen based on their strike rate with bowlers. Cricket like many other games is a game of strategy, skill, talent and some amount of luck. So while the LP formulation can provide some direction,  one must be aware anything could happen in a game of cricket!

Thoughts, comments, suggestions welcome!

Also see
1. Inswinger: yorkr swings into International T20s
2. Working with Node.js and PostgreSQL
3. Simulating the domino effect in Android using Box2D and AndEngine
4. Introducing cricket package yorkr: Part 1- Beaten by sheer pace!
5. Introducing QCSimulator: A 5-qubit quantum computing simulator in R
6. A Cloud medley with IBM Bluemix, Cloudant DB and Node.js

To see all posts see Index of Posts

cricketr flexes new muscles: The final analysis


Twas brillig, and the slithy toves
Did gyre and gimble in the wabe:
All mimsy were the borogoves,
And the mome raths outgrabe.

       Jabberwocky by Lewis Carroll
                   

No analysis of cricket is complete, without determining how players would perform in the host country. Playing Test cricket on foreign pitches, in the host country, is a ‘real test’ for both batsmen and bowlers. Players, who can perform consistently both on domestic and foreign pitches are the genuinely ‘class’ players. Player performance on foreign pitches lets us differentiate the paper tigers, and home ground bullies among batsmen. Similarly, spinners who perform well, only on rank turners in home ground or pace bowlers who can only swing and generate bounce on specially prepared pitches are neither  genuine spinners nor  real pace bowlers.

So this post, helps in identifying those with real strengths, and those who play good only when the conditions are in favor, in home grounds. This post brings a certain level of finality to the analysis of players with my R package ‘cricketr’

Besides, I also meant ‘final analysis’ in the literal sense, as I intend to take a long break from cricket analysis/analytics and focus on some other domains like Neural Networks, Deep Learning and Spark.

As already mentioned, my R package ‘cricketr’ uses the statistics info available in ESPN Cricinfo Statsguru. You should be able to install the package from CRAN and use many of the functions available in the package. Please be mindful of ESPN Cricinfo Terms of Use

(Note: This page is also hosted at RPubs as cricketrFinalAnalysis. You can download the PDF file at cricketrFinalAnalysis.

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

For getting data of a player against a particular country for the match played in the host country, I just had to add 2 extra parameters to the getPlayerData() function. The cricketr package has been updated with the changed functions for getPlayerData() – Tests, getPlayerDataOD() – ODI and getPlayerDataTT() for the Twenty20s. The updated functions will be available in cricketr Version -0.0.14

The data for the following players have already been obtained with the new, changed getPlayerData() function and have been saved as *.csv files. I will be re-using these files, instead of getting them all over again. Hence the getPlayerData() lines have been commented below

library(cricketr)

1. Performance of a batsman against a host ountry in the host country

For e.g We can the get the data for Sachin Tendulkar for matches played against Australia and in Australia Here opposition=2 and host =2 indicate that the opposition is Australia and the host country is also Australia

#tendulkarAus=getPlayerData(35320,opposition=2,host=2,file="tendulkarVsAusInAus.csv",type="batting")

All cricketr functions can be used with this data frame, as before. All the charts show the performance of Tendulkar in Australia against Australia.

par(mfrow=c(2,3))
par(mar=c(4,4,2,2))
batsman4s("./data/tendulkarVsAusInAus.csv","Tendulkar")
batsman6s("./data/tendulkarVsAusInAus.csv","Tendulkar")
batsmanRunsRanges("./data/tendulkarVsAusInAus.csv","Tendulkar")
batsmanDismissals("./data/tendulkarVsAusInAus.csv","Tendulkar")
batsmanAvgRunsGround("./data/tendulkarVsAusInAus.csv","Tendulkar")
batsmanMovingAverage("./data/tendulkarVsAusInAus.csv","Tendulkar")

dev.off()
## null device 
##           1

2. Relative performances of international batsmen against England in England

While we can analyze the performance of a player against an opposition in some host country, I wanted to compare the relative performances of players, to see how players from different nations play in a host country which is not their home ground.

The following lines gets player’s data of matches played in England and against England.The Oval, Lord’s are famous for generating some dangerous swing and bounce. I chose the following players

  1. Sir Don Bradman (Australia)
  2. Steve Waugh (Australia)
  3. Rahul Dravid (India)
  4. Vivian Richards (West Indies)
  5. Sachin Tendulkar (India)
#tendulkarEng=getPlayerData(35320,opposition=1,host=1,file="tendulkarVsEngInEng.csv",type="batting")
#bradmanEng=getPlayerData(4188,opposition=1,host=1,file="bradmanVsEngInEng.csv",type="batting")
#srwaughEng=getPlayerData(8192,opposition=1,host=1,file="srwaughVsEngInEng.csv",type="batting")
#dravidEng=getPlayerData(28114,opposition=1,host=1,file="dravidVsEngInEng.csv",type="batting")
#vrichardEng=getPlayerData(52812,opposition=1,host=1,file="vrichardsEngInEng.csv",type="batting")
frames <- list("./data/tendulkarVsEngInEng.csv","./data/bradmanVsEngInEng.csv","./data/srwaughVsEngInEng.csv",
               "./data/dravidVsEngInEng.csv","./data/vrichardsEngInEng.csv")
names <- list("S Tendulkar","D Bradman","SR Waugh","R Dravid","Viv Richards")

The Lords and the Oval in England are some of the best pitches in the world. Scoring on these pitches and weather conditions, where there is both swing and bounce really requires excellent batting skills. It can be easily seen that Don Bradman stands heads and shoulders over everybody else, averaging close a cumulative average of 100+. He is followed by Viv Richards, who averages around ~60. Interestingly in English conditions, Rahul Dravid edges out Sachin Tendulkar.

relativeBatsmanCumulativeAvgRuns(frames,names)

# The other 2 plots on relative strike rate and cumulative average strike rate,
shows Viv Richards really  blasts the bowling. Viv Richards has a strike rate 
of 70, while Bradman 62+, followed by Tendulkar.
relativeBatsmanSR(frames,names)

relativeBatsmanCumulativeStrikeRate(frames,names)

3. Relative performances of international batsmen against Australia in Australia

The following players from these countries were chosen

  1. Sachin Tendulkar (India)
  2. Viv Richard (West Indies)
  3. David Gower (England)
  4. Jacques Kallis (South Africa)
  5. Alastair Cook (Emgland)
frames <- list("./data/tendulkarVsAusInAus.csv","./data/vrichardsVAusInAus.csv","./data/dgowerVsAusInAus.csv",
               "./data/kallisVsAusInAus.csv","./data/ancookVsWIInWI.csv")
names <- list("S Tendulkar","Viv Richards","David Gower","J Kallis","AN Cook")

Alastair Cook of England has fantastic cumulative average of 55+ on the pitches of Australia. There is a dip towards the end, but we cannot predict whether it would have continued. AN Cook is followed by Tendulkar who has a steady average of 50+ runs, after which there is Viv Richards.

relativeBatsmanCumulativeAvgRuns(frames,names)

#With respect to cumulative or relative strike rate Viv Richards is a class apart.He seems to really
#tear into bowlers. David Gower has an excellent strike rate and is followed by Tendulkar
relativeBatsmanSR(frames,names)

relativeBatsmanCumulativeStrikeRate(frames,names)

4. Relative performances of international batsmen against India in India

While England & Australia are famous for bouncy tracks with swing, Indian pitches are renowed for being extraordinary turners. Also India has always thrown up world class spinners, from the spin quartet of BS Chandraskehar, Bishen Singh Bedi, EAS Prasanna, S Venkatraghavan, to the times of dangerous Anil Kumble, and now to the more recent Ravichander Ashwon and Harbhajan Singh.

A batsmen who can score runs in India against Indian spinners has to be really adept in handling all kinds of spin.

While Clive Lloyd & Alvin Kallicharan had the best performance against India, they have not been included as ESPN Cricinfo had many of the columns missing.

So I chose the following international players for the analysis against India

  1. Hashim Amla (South Africa)
  2. Alastair Cook (England)
  3. Matthew Hayden (Australia)
  4. Viv Richards (West Indies)
frames <- list("./data/amlaVsIndInInd.csv","./data/ancookVsIndInInd.csv","./data/mhaydenVsIndInInd.csv",
               "./data/vrichardsVsIndInInd.csv")
names <- list("H Amla","AN Cook","M Hayden","Viv Riachards")

Excluding Clive Lloyd & Alvin Kallicharan the next best performer against India is Hashim Amla,followed by Alastair Cook, Viv Richards.

relativeBatsmanCumulativeAvgRuns(frames,names)

#With respect to strike rate, there is no contest when Viv Richards is around. He is clearly the best 
#striker of the ball regardless of whether it is the pacy wickets of 
#Australia/England or the spinning tracks of the subcontinent. After 
#Viv Richards, Hayden and Alastair Cook have good cumulative strike rates
#in India
relativeBatsmanSR(frames,names)

relativeBatsmanCumulativeStrikeRate(frames,names)

5. All time greats of Indian batting

I couldn’t resist checking out how the top Indian batsmen perform when playing in host countries So here is a look at how the top Indian batsmen perform against different host countries

6. Top Indian batsmen against Australia in Australia

The following Indian batsmen were chosen

  1. Sunil Gavaskar
  2. Sachin Tendulkar
  3. Virat Kohli
  4. Virendar Sehwag
  5. VVS Laxman
frames <- list("./data/tendulkarVsAusInAus.csv","./data/gavaskarVsAusInAus.csv","./data/kohliVsAusInAus.csv",
               "./data/sehwagVsAusInAus.csv","./data/vvslaxmanVsAusInAus.csv")
names <- list("S Tendulkar","S Gavaskar","V Kohli","V Sehwag","VVS Laxman")

Virat Kohli has the best overall performance against Australia, with a current cumulative average of 60+ runs for the total number of innings played by him (15). With 15 matches the 2nd best is Virendar Sehwag, followed by VVS Laxman. Tendulkar maintains a cumulative average of 48+ runs for an excess of 30+ innings.

relativeBatsmanCumulativeAvgRuns(frames,names)

# Sehwag leads the strike rate against host Australia, followed by 
# Tendulkar in Australia and then Kohli
relativeBatsmanSR(frames,names)

relativeBatsmanCumulativeStrikeRate(frames,names)

7. Top Indian batsmen against England in England

The top Indian batmen’s performances against England are shown below

  1. Rahul Dravid
  2. Dilip Vengsarkar
  3. Rahul Dravid
  4. Sourav Ganguly
  5. Virat Kohli
frames <- list("./data/tendulkarVsEngInEng.csv","./data/dravidVsEngInEng.csv","./data/vengsarkarVsEngInEng.csv",
               "./data/gangulyVsEngInEng.csv","./data/gavaskarVsEngInEng.csv","./data/kohliVsEngInEng.csv")
names <- list("S Tendulkar","R Dravid","D Vengsarkar","S Ganguly","S Gavaskar","V Kohli")

Rahul Dravid has the best performance against England and edges out Tendulkar. He is followed by Tendulkar and then Sourav Ganguly. Note:Incidentally Virat Kohli’s performance against England in England so far has been extremely poor and he averages around 13-15 runs per innings. However he has a long way to go and I hope he catches up. In any case it will be an uphill climb for Kohli in England.

relativeBatsmanCumulativeAvgRuns(frames,names)

#Tendulkar, Ganguly and Dravid have the best strike rate and in that order.
relativeBatsmanSR(frames,names)

relativeBatsmanCumulativeStrikeRate(frames,names)

8. Top Indian batsmen against West Indies in West Indies

frames <- list("./data/tendulkarVsWInWI.csv","./data/dravidVsWInWI.csv","./data/vvslaxmanVsWIInWI.csv",
               "./data/gavaskarVsWIInWI.csv")
names <- list("S Tendulkar","R Dravid","VVS Laxman","S Gavaskar")

Against the West Indies Sunil Gavaskar is heads and shoulders above the rest. Gavaskar has a very impressive cumulative average against West Indies

relativeBatsmanCumulativeAvgRuns(frames,names)

# VVS Laxman followed by  Tendulkar & then Dravid have a very 
# good strike rate against the West Indies
relativeBatsmanCumulativeStrikeRate(frames,names)

9. World’s best spinners on tracks suited for pace & bounce

In this part I compare the performances of the top 3 spinners in recent years and check out how they perform on surfaces that are known for pace, and bounce. I have taken the following 3 spinners

  1. Anil Kumble (India)
  2. M Muralitharan (Sri Lanka)
  3. Shane Warne (Australia)
#kumbleEng=getPlayerData(30176  ,opposition=3,host=3,file="kumbleVsEngInEng.csv",type="bowling")
#muraliEng=getPlayerData(49636  ,opposition=3,host=3,file="muraliVsEngInEng.csv",type="bowling")
#warneEng=getPlayerData(8166  ,opposition=3,host=3,file="warneVsEngInEng.csv",type="bowling")

10. Top international spinners against England in England

frames <- list("./data/kumbleVsEngInEng.csv","./data/muraliVsEngInEng.csv","./data/warneVsEngInEng.csv")
names <- list("Anil KUmble","M Muralitharan","Shane Warne")

Against England and in England, Muralitharan shines with a cumulative average of nearly 5 wickets per match with a peak of almost 8 wickets. Shane Warne has a steady average at 5 wickets and then Anil Kumble.

relativeBowlerCumulativeAvgWickets(frames,names)

# The order relative cumulative Economy rate, Warne has the best figures,followed by Anil Kumble. Muralitharan
# is much more expensive.
relativeBowlerCumulativeAvgEconRate(frames,names)

11. Top international spinners against South Africa in South Africa

frames <- list("./data/kumbleVsSAInSA.csv","./data/muraliVsSAInSA.csv","./data/warneVsSAInSA.csv")
names <- list("Anil Kumble","M Muralitharan","Shane Warne")

In South Africa too, Muralitharan has the best wicket taking performance averaging about 4 wickets. Warne averages around 3 wickets and Kumble around 2 wickets

relativeBowlerCumulativeAvgWickets(frames,names)

# Muralitharan is expensive in South Africa too, while Kumble and Warne go neck-to-neck in the economy rate.
# Kumble edges out Warne and has a better cumulative average economy rate
relativeBowlerCumulativeAvgEconRate(frames,names)

11. Top international pacers against India in India

As a final analysis I check how the world’s pacers perform in India against India. India pitches are supposed to be flat devoid of bounce, while being terrific turners. Hence Indian pitches are more suited to spin bowling than pace bowling. This is changing these days.

The best performers against India in India are mostly the deadly pacemen of yesteryears

For this I have chosen the following bowlers

  1. Courtney Walsh (West Indies)
  2. Andy Roberts (West Indies)
  3. Malcolm Marshall
  4. Glenn McGrath
#cawalshInd=getPlayerData(53216  ,opposition=6,host=6,file="cawalshVsIndInInd.csv",type="bowling")
#arobertsInd=getPlayerData(52817  ,opposition=6,host=6,file="arobertsIndInInd.csv",type="bowling")
#mmarshallInd=getPlayerData(52419  ,opposition=6,host=6,file="mmarshallVsIndInInd.csv",type="bowling")
#gmccgrathInd=getPlayerData(6565  ,opposition=6,host=6,file="mccgrathVsIndInInd.csv",type="bowling")
frames <- list("./data/cawalshVsIndInInd.csv","./data/arobertsIndInInd.csv","./data/mmarshallVsIndInInd.csv",
               "./data/mccgrathVsIndInInd.csv")
names <- list("C Walsh","A Roberts","M Marshall","G McGrath")

Courtney Walsh has the best performance, followed by Andy Roberts followed by Andy Roberts and then Malcom Marshall who tips ahead of Glenn McGrath

relativeBowlerCumulativeAvgWickets(frames,names)

#On the other hand McGrath has the best economy rate, followed by A Roberts and then Courtney Walsh
relativeBowlerCumulativeAvgEconRate(frames,names)

12. ODI performance of a player against a specific country in the host country

This gets the data for MS Dhoni in ODI matches against Australia and in Australia

#dhoniAusODI=getPlayerDataOD(28081,opposition=2,host=2,file="dhoniVsAusInAusODI.csv",type="batting")

13. Twenty 20 performance of a player against a specific country in the host country

#dhoniAusTT=getPlayerDataOD(28081,opposition=2,host=2,file="dhoniVsAusInAusTT.csv",type="batting")

All the ODI and Twenty20 functions of cricketr can be used on the above dataframes of MS Dhoni.

Some key observations

Here are some key observations

  1. At the top of the batting spectrum is Don Bradman with a very impressive average 100-120 in matches played in England and Australia. Unfortunately there weren’t matches he played in other countries and different pitches. 2.Viv Richard has the best cumulative strike rate overall.
  2. Muralitharan strikes more often than Kumble or Warne even in pitches at ENgland, South Africa and West Indies. However Muralitharan is also the most expensive
  3. Warne and Kumble have a much better economy rate than Muralitharan.
  4. Sunil Gavaskar has an extremely impressive performance in West Indies.
  5. Rahul Dravid performs much better than Tendulkar in both England and West Indies.
  6. Virat Kohli has the best performance against Australia so far and hope he maintains his stellar performance followed by Sehwag. However Kohli’s performance in England has been very poor
  7. West Indies batsmen and bowlers seem to thrive on Indian pitches, with Clive Lloyd and Alvin Kalicharan at the top of the list.

You may like my Shiny apps on cricket

  1. Inswinger- Analyzing International. T20s
  2. GooglyPlus – An app for analyzing IPL
  3. Sixer – App based on R package cricketr

Also see

  1. Exploring Quantum Gate operations with QCSimulator
  2. Neural Networks: The mechanics of backpropagation
  3. Re-introducing cricketr! : An R package to analyze performances of cricketers
  4. yorkr crashes the IPL party ! – Part 1
  5. cricketr and yorkr books – Paperback now in Amazon
  6.  Hand detection through Haartraining: A hands-on approach

To see all my posts see Index of posts

Analysis of IPL T20 matches with yorkr templates


Introduction

In this post I create RMarkdown templates for end-to-end analysis of IPL T20 matches, that are available on Cricsheet based on my R package yorkr.  With these templates you can convert all IPL data which is in yaml format to R dataframes. Further I create data and the necessary templates for analyzing IPL matches, teams and players. All of these can be accessed at yorkrIPLTemplate.

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

The templates are

  1. Template for conversion and setup – IPLT20Template.Rmd
  2. Any IPL match – IPLMatchtemplate.Rmd
  3. IPL matches between 2 nations – IPLMatches2TeamTemplate.Rmd
  4. A IPL nations performance against all other IPL nations – IPLAllMatchesAllOppnTemplate.Rmd
  5. Analysis of IPL batsmen and bowlers of all IPL nations – IPLBatsmanBowlerTemplate.Rmd

Besides the templates the repository also includes the converted data for all IPL matches I downloaded from Cricsheet in Dec 2016. So this data is complete till the 2016 IPL season. You can recreate the files as more matches are added to Cricsheet site in IPL 2017 and future seasons. This post contains all the steps needed for detailed analysis of IPL matches, teams and IPL player. This will also be my reference in future if I decide to analyze IPL in future!

There will be 5 folders at the root

  1. IPLdata – Match files as yaml from Cricsheet
  2. IPLMatches – Yaml match files converted to dataframes
  3. IPLMatchesBetween2Teams – All Matches between any 2 IPL teams
  4. allMatchesAllOpposition – An IPL teams’s performance against all other teams
  5. BattingBowlingDetails – Batting and bowling details of all IPL teams
library(yorkr)
library(dplyr)

The first few steps take care of the data setup. This needs to be done before any of the analysis of IPL batsmen, bowlers, any IPL match, matches between any 2 IPL countries or analysis of a teams performance against all other countries

There will be 5 folders at the root

  1. data
  2. IPLMatches
  3. IPLMatchesBetween2Teams
  4. allMatchesAllOpposition
  5. BattingBowlingDetails

The source YAML files will be in IPLData folder

1.Create directory of IPLMatches

Some files may give conversions errors. You could try to debug the problem or just remove it from the IPLdata folder. At most 2-4 file will have conversion problems and I usally remove then from the files to be converted.

Also take a look at my GooglyPlus shiny app which was created after performing the same conversion on the Dec 16 data .

convertAllYaml2RDataframesT20("data","IPLMatches")

2.Save all matches between all combinations of IPL nations

This function will create the set of all matches between each IPL team against every other IPL team. This uses the data that was created in IPLMatches, with the convertAllYaml2RDataframesIPL() function.

setwd("./IPLMatchesBetween2Teams")
saveAllMatchesBetween2IPLTeams("../IPLMatches")

3.Save all matches against all opposition

This will create a consolidated dataframe of all matches played by every IPL playing nation against all other nattions. This also uses the data that was created in IPLMatches, with the convertAllYaml2RDataframesIPL() function.

setwd("../allMatchesAllOpposition")
saveAllMatchesAllOppositionIPLT20("../IPLMatches")

4. Create batting and bowling details for each IPL team

These are the current IPL playing teams. You can add to this vector as newer IPL teams start playing IPL. You will get to know all IPL teams by also look at the directory created above namely allMatchesAllOpposition. This also uses the data that was created in IPLMatches, with the convertAllYaml2RDataframesIPL() function.

setwd("../BattingBowlingDetails")
ipl_teams <- list("Chennai Super Kings","Deccan Chargers", "Delhi Daredevils","Kings XI Punjab", 
              "Kochi Tuskers Kerala","Kolkata Knight Riders","Mumbai Indians","Pune Warriors",
              "Rajasthan Royals","Royal Challengers Bangalore","Sunrisers Hyderabad","Gujarat Lions",
                 "Rising Pune Supergiants")

for(i in seq_along(ipl_teams)){
    print(ipl_teams[i])
    val <- paste(ipl_teams[i],"-details",sep="")
    val <- getTeamBattingDetails(ipl_teams[i],dir="../IPLMatches", save=TRUE)

}

for(i in seq_along(ipl_teams)){
    print(ipl_teams[i])
    val <- paste(ipl_teams[i],"-details",sep="")
    val <- getTeamBowlingDetails(ipl_teams[i],dir="../IPLMatches", save=TRUE)

}

5. Get the list of batsmen for a particular IPL team

The following code is needed for analyzing individual IPL batsmen. In IPL a player could have played in multiple IPL teams.

getBatsmen <- function(df){
    bmen <- df %>% distinct(batsman) 
    bmen <- as.character(bmen$batsman)
    batsmen <- sort(bmen)
}
load("Chennai Super Kings-BattingDetails.RData")
csk_details <- battingDetails
load("Deccan Chargers-BattingDetails.RData")
dc_details <- battingDetails
load("Delhi Daredevils-BattingDetails.RData")
dd_details <- battingDetails
load("Kings XI Punjab-BattingDetails.RData")
kxip_details <- battingDetails
load("Kochi Tuskers Kerala-BattingDetails.RData")
ktk_details <- battingDetails
load("Kolkata Knight Riders-BattingDetails.RData")
kkr_details <- battingDetails
load("Mumbai Indians-BattingDetails.RData")
mi_details <- battingDetails
load("Pune Warriors-BattingDetails.RData")
pw_details <- battingDetails
load("Rajasthan Royals-BattingDetails.RData")
rr_details <- battingDetails
load("Royal Challengers Bangalore-BattingDetails.RData")
rcb_details <- battingDetails
load("Sunrisers Hyderabad-BattingDetails.RData")
sh_details <- battingDetails
load("Gujarat Lions-BattingDetails.RData")
gl_details <- battingDetails
load("Rising Pune Supergiants-BattingDetails.RData")
rps_details <- battingDetails

#Get the batsmen for each IPL team
csk_batsmen <- getBatsmen(csk_details)
dc_batsmen <- getBatsmen(dc_details)
dd_batsmen <- getBatsmen(dd_details)
kxip_batsmen <- getBatsmen(kxip_details)
ktk_batsmen <- getBatsmen(ktk_details)
kkr_batsmen <- getBatsmen(kkr_details)
mi_batsmen <- getBatsmen(mi_details)
pw_batsmen <- getBatsmen(pw_details)
rr_batsmen <- getBatsmen(rr_details)
rcb_batsmen <- getBatsmen(rcb_details)
sh_batsmen <- getBatsmen(sh_details)
gl_batsmen <- getBatsmen(gl_details)
rps_batsmen <- getBatsmen(rps_details)

# Save the dataframes
save(csk_batsmen,file="csk.RData")
save(dc_batsmen, file="dc.RData")
save(dd_batsmen, file="dd.RData")
save(kxip_batsmen, file="kxip.RData")
save(ktk_batsmen, file="ktk.RData")
save(kkr_batsmen, file="kkr.RData")
save(mi_batsmen , file="mi.RData")
save(pw_batsmen, file="pw.RData")
save(rr_batsmen, file="rr.RData")
save(rcb_batsmen, file="rcb.RData")
save(sh_batsmen, file="sh.RData")
save(gl_batsmen, file="gl.RData")
save(rps_batsmen, file="rps.RData")

6. Get the list of bowlers for a particular IPL team

The method below can get the list of bowler names for any IPL team.The following code is needed for analyzing individual IPL bowlers. In IPL a player could have played in multiple IPL teams.

getBowlers <- function(df){
    bwlr <- df %>% distinct(bowler) 
    bwlr <- as.character(bwlr$bowler)
    bowler <- sort(bwlr)
}

load("Chennai Super Kings-BowlingDetails.RData")
csk_details <- bowlingDetails
load("Deccan Chargers-BowlingDetails.RData")
dc_details <- bowlingDetails
load("Delhi Daredevils-BowlingDetails.RData")
dd_details <- bowlingDetails
load("Kings XI Punjab-BowlingDetails.RData")
kxip_details <- bowlingDetails
load("Kochi Tuskers Kerala-BowlingDetails.RData")
ktk_details <- bowlingDetails
load("Kolkata Knight Riders-BowlingDetails.RData")
kkr_details <- bowlingDetails
load("Mumbai Indians-BowlingDetails.RData")
mi_details <- bowlingDetails
load("Pune Warriors-BowlingDetails.RData")
pw_details <- bowlingDetails
load("Rajasthan Royals-BowlingDetails.RData")
rr_details <- bowlingDetails
load("Royal Challengers Bangalore-BowlingDetails.RData")
rcb_details <- bowlingDetails
load("Sunrisers Hyderabad-BowlingDetails.RData")
sh_details <- bowlingDetails
load("Gujarat Lions-BowlingDetails.RData")
gl_details <- bowlingDetails
load("Rising Pune Supergiants-BowlingDetails.RData")
rps_details <- bowlingDetails

# Get the bowlers for each team
csk_bowlers <- getBowlers(csk_details)
dc_bowlers <- getBowlers(dc_details)
dd_bowlers <- getBowlers(dd_details)
kxip_bowlers <- getBowlers(kxip_details)
ktk_bowlers <- getBowlers(ktk_details)
kkr_bowlers <- getBowlers(kkr_details)
mi_bowlers <- getBowlers(mi_details)
pw_bowlers <- getBowlers(pw_details)
rr_bowlers <- getBowlers(rr_details)
rcb_bowlers <- getBowlers(rcb_details)
sh_bowlers <- getBowlers(sh_details)
gl_bowlers <- getBowlers(gl_details)
rps_bowlers <- getBowlers(rps_details)

#Save the dataframes
save(csk_bowlers,file="csk1.RData")
save(dc_bowlers, file="dc1.RData")
save(dd_bowlers, file="dd1.RData")
save(kxip_bowlers, file="kxip1.RData")
save(ktk_bowlers, file="ktk1.RData")
save(kkr_bowlers, file="kkr1.RData")
save(mi_bowlers , file="mi1.RData")
save(pw_bowlers, file="pw1.RData")
save(rr_bowlers, file="rr1.RData")
save(rcb_bowlers, file="rcb1.RData")
save(sh_bowlers, file="sh1.RData")
save(gl_bowlers, file="gl1.RData")
save(rps_bowlers, file="rps1.RData")

Now we are all set

A)  IPL T20 Match Analysis

1 IPL Match Analysis

Load any match data from the ./IPLMatches folder for e.g. Chennai Super Kings-Deccan Chargers-2008-05-06.RData

setwd("./IPLMatches")
load("Chennai Super Kings-Deccan Chargers-2008-05-06.RData")
csk_dc<- overs
#The steps are
load("IPLTeam1-IPLTeam2-Date.Rdata")
IPLTeam1_IPLTeam2 <- overs

All analysis for this match can be done now

2. Scorecard

teamBattingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam1")
teamBattingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam2")

3.Batting Partnerships

teamBatsmenPartnershipMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
teamBatsmenPartnershipMatch(IPLTeam1_IPLTeam2,"IPLTeam2","IPLTeam1")

4. Batsmen vs Bowler Plot

teamBatsmenVsBowlersMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=TRUE)
teamBatsmenVsBowlersMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)

5. Team bowling scorecard

teamBowlingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam1")
teamBowlingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam2")

6. Team bowling Wicket kind match

teamBowlingWicketKindMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <-teamBowlingWicketKindMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

7. Team Bowling Wicket Runs Match

teamBowlingWicketRunsMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <-teamBowlingWicketRunsMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

8. Team Bowling Wicket Match

m <-teamBowlingWicketMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m
teamBowlingWicketMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")

9. Team Bowler vs Batsmen

teamBowlersVsBatsmenMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <- teamBowlersVsBatsmenMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

10. Match Worm chart

matchWormGraph(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")

B)  IPL  Matches between 2  IPL teams

1 IPL Match Analysis

Load any match data from the ./IPLMatches folder for e.g. Chennai Super Kings-Deccan Chargers-2008-05-06.RData

setwd("./IPLMatches")
load("Chennai Super Kings-Deccan Chargers-2008-05-06.RData")
csk_dc<- overs
#The steps are
load("IPLTeam1-IPLTeam2-Date.Rdata")
IPLTeam1_IPLTeam2 <- overs

All analysis for this match can be done now

2. Scorecard

teamBattingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam1")
teamBattingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam2")

3.Batting Partnerships

teamBatsmenPartnershipMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
teamBatsmenPartnershipMatch(IPLTeam1_IPLTeam2,"IPLTeam2","IPLTeam1")

4. Batsmen vs Bowler Plot

teamBatsmenVsBowlersMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=TRUE)
teamBatsmenVsBowlersMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)

5. Team bowling scorecard

teamBowlingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam1")
teamBowlingScorecardMatch(IPLTeam1_IPLTeam2,"IPLTeam2")

6. Team bowling Wicket kind match

teamBowlingWicketKindMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <-teamBowlingWicketKindMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

7. Team Bowling Wicket Runs Match

teamBowlingWicketRunsMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <-teamBowlingWicketRunsMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

8. Team Bowling Wicket Match

m <-teamBowlingWicketMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m
teamBowlingWicketMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")

9. Team Bowler vs Batsmen

teamBowlersVsBatsmenMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")
m <- teamBowlersVsBatsmenMatch(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2",plot=FALSE)
m

10. Match Worm chart

matchWormGraph(IPLTeam1_IPLTeam2,"IPLTeam1","IPLTeam2")

C)  IPL Matches for a team against all other teams

1. IPL Matches for a team against all other teams

Load the data between for a IPL team against all other countries ./allMatchesAllOpposition for e.g all matches of Kolkata Knight Riders

load("allMatchesAllOpposition-Kolkata Knight Riders.RData")
kkr_matches <- matches
IPLTeam="IPLTeam1"
allMatches <- paste("allMatchesAllOposition-",IPLTeam,".RData",sep="")
load(allMatches)
IPLTeam1AllMatches <- matches

2. Team’s batting scorecard all Matches

m <-teamBattingScorecardAllOppnAllMatches(IPLTeam1AllMatches,theTeam="IPLTeam1")
m

3. Batting scorecard of opposing team

m <-teamBattingScorecardAllOppnAllMatches(matches=IPLTeam1AllMatches,theTeam="IPLTeam2")

4. Team batting partnerships

m <- teamBatsmenPartnershipAllOppnAllMatches(IPLTeam1AllMatches,theTeam="IPLTeam1")
m
m <- teamBatsmenPartnershipAllOppnAllMatches(IPLTeam1AllMatches,theTeam='IPLTeam1',report="detailed")
head(m,30)
m <- teamBatsmenPartnershipAllOppnAllMatches(IPLTeam1AllMatches,theTeam='IPLTeam1',report="summary")
m

5. Team batting partnerships plot

teamBatsmenPartnershipAllOppnAllMatchesPlot(IPLTeam1AllMatches,"IPLTeam1",main="IPLTeam1")
teamBatsmenPartnershipAllOppnAllMatchesPlot(IPLTeam1AllMatches,"IPLTeam1",main="IPLTeam2")

6, Team batsmen vs bowlers report

m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(IPLTeam1AllMatches,"IPLTeam1",rank=0)
m
m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(IPLTeam1AllMatches,"IPLTeam1",rank=1,dispRows=30)
m
m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(matches=IPLTeam1AllMatches,theTeam="IPLTeam2",rank=1,dispRows=25)
m

7. Team batsmen vs bowler plot

d <- teamBatsmenVsBowlersAllOppnAllMatchesRept(IPLTeam1AllMatches,"IPLTeam1",rank=1,dispRows=50)
d
teamBatsmenVsBowlersAllOppnAllMatchesPlot(d)
d <- teamBatsmenVsBowlersAllOppnAllMatchesRept(IPLTeam1AllMatches,"IPLTeam1",rank=2,dispRows=50)
teamBatsmenVsBowlersAllOppnAllMatchesPlot(d)

8. Team bowling scorecard

teamBowlingScorecardAllOppnAllMatchesMain(matches=IPLTeam1AllMatches,theTeam="IPLTeam1")
teamBowlingScorecardAllOppnAllMatches(IPLTeam1AllMatches,'IPLTeam2')

9. Team bowler vs batsmen

teamBowlersVsBatsmenAllOppnAllMatchesMain(IPLTeam1AllMatches,theTeam="IPLTeam1",rank=0)
teamBowlersVsBatsmenAllOppnAllMatchesMain(IPLTeam1AllMatches,theTeam="IPLTeam1",rank=2)
teamBowlersVsBatsmenAllOppnAllMatchesRept(matches=IPLTeam1AllMatches,theTeam="IPLTeam1",rank=0)

10. Team Bowler vs bastmen

df <- teamBowlersVsBatsmenAllOppnAllMatchesRept(IPLTeam1AllMatches,theTeam="IPLTeam1",rank=1)
teamBowlersVsBatsmenAllOppnAllMatchesPlot(df,"IPLTeam1","IPLTeam1")

11. Team bowler wicket kind

teamBowlingWicketKindAllOppnAllMatches(IPLTeam1AllMatches,t1="IPLTeam1",t2="All")
teamBowlingWicketKindAllOppnAllMatches(IPLTeam1AllMatches,t1="IPLTeam1",t2="IPLTeam2")

12.

teamBowlingWicketRunsAllOppnAllMatches(IPLTeam1AllMatches,t1="IPLTeam1",t2="All",plot=TRUE)
teamBowlingWicketRunsAllOppnAllMatches(IPLTeam1AllMatches,t1="IPLTeam1",t2="IPLTeam2",plot=TRUE)

1 IPL Batsman setup functions

Get the batsman’s details for a batsman

setwd("../BattingBowlingDetails")
# IPL Team names
IPLTeamNames <- list("Chennai Super Kings","Deccan Chargers", "Delhi Daredevils","Kings Xi Punjab", 
                  "Kochi Tuskers Kerala","Kolkata Knight Riders","Mumbai Indians","Pune Warriors",
                  "Rajasthan Royals","Royal Challengers Bangalore","Sunrisers Hyderabad","Gujarat Lions",
                  "Rising Pune Supergiants")           


# Check and get the team indices of IPL teams in which the batsman has played
getTeamIndex <- function(batsman){
    setwd("./BattingBowlingDetails")
    load("csk.RData")
    load("dc.RData")
    load("dd.RData")
    load("kxip.RData")
    load("ktk.RData")
    load("kkr.RData")
    load("mi.RData")
    load("pw.RData")
    load("rr.RData")
    load("rcb.RData")
    load("sh.RData")
    load("gl.RData")
    load("rps.RData")
    setwd("..")
    getwd()
    print(ls())
    teams_batsmen = list(csk_batsmen,dc_batsmen,dd_batsmen,kxip_batsmen,ktk_batsmen,kkr_batsmen,mi_batsmen,
                         pw_batsmen,rr_batsmen,rcb_batsmen,sh_batsmen,gl_batsmen,rps_batsmen)
    b <- NULL
    for (i in 1:length(teams_batsmen)){
        a <- which(teams_batsmen[[i]] == batsman)

        if(length(a) != 0)
            b <- c(b,i)
    }
    b
}

# Get the list of the IPL team names from the indices passed
getTeams <- function(x){

    l <- NULL
    # Get the teams passed in as indexes
    for (i in seq_along(x)){

        l <- c(l, IPLTeamNames[[x[i]]]) 

    }
    l
}

# Create a consolidated data frame with all teams the IPL batsman has played for
getIPLBatsmanDF <- function(teamNames){
    batsmanDF <- NULL
   # Create a consolidated Data frame of batsman for all IPL teams played
    for (i in seq_along(teamNames)){
       df <- getBatsmanDetails(team=teamNames[i],name=IPLBatsman,dir="./BattingBowlingDetails")
       batsmanDF <- rbind(batsmanDF,df) 

    }
    batsmanDF
}

2. Create a consolidated IPL batsman data frame

# Since an IPL batsman coculd have played in multiple teams we need to determine these teams and
# create a consolidated data frame for the analysis
# For example to check MS Dhoni we need to do the following

IPLBatsman = "MS Dhoni"
#Check and get the team indices of IPL teams in which the batsman has played
i <- getTeamIndex(IPLBatsman)

# Get the team names in which the IPL batsman has played
teamNames <- getTeams(i)
    # Check if file exists in the directory. This check is necessary when moving between matchType


############## Create a consolidated IPL batsman dataframe for analysis
batsmanDF <- getIPLBatsmanDF(teamNames)

3. Runs vs deliveries

# For e.g. batsmanName="MS Dhoni""
#batsmanRunsVsDeliveries(batsmanDF, "MS Dhoni")
batsmanRunsVsDeliveries(batsmanDF,"batsmanName")

4. Batsman 4s & 6s

batsman46 <- select(batsmanDF,batsman,ballsPlayed,fours,sixes,runs)
p1 <- batsmanFoursSixes(batsman46,"batsmanName")

5. Batsman dismissals

batsmanDismissals(batsmanDF,"batsmanName")

6. Runs vs Strike rate

batsmanRunsVsStrikeRate(batsmanDF,"batsmanName")

7. Batsman Moving Average

batsmanMovingAverage(batsmanDF,"batsmanName")

8. Batsman cumulative average

batsmanCumulativeAverageRuns(batsmanDF,"batsmanName")

9. Batsman cumulative strike rate

batsmanCumulativeStrikeRate(batsmanDF,"batsmanName")

10. Batsman runs against oppositions

batsmanRunsAgainstOpposition(batsmanDF,"batsmanName")

11. Batsman runs vs venue

batsmanRunsVenue(batsmanDF,"batsmanName")

12. Batsman runs predict

batsmanRunsPredict(batsmanDF,"batsmanName")

13.Bowler set up functions

setwd("../BattingBowlingDetails")
# IPL Team names
IPLTeamNames <- list("Chennai Super Kings","Deccan Chargers", "Delhi Daredevils","Kings Xi Punjab", 
                  "Kochi Tuskers Kerala","Kolkata Knight Riders","Mumbai Indians","Pune Warriors",
                  "Rajasthan Royals","Royal Challengers Bangalore","Sunrisers Hyderabad","Gujarat Lions",
                  "Rising Pune Supergiants")    



# Get the team indices of IPL teams for which the bowler as played
getTeamIndex_bowler <- function(bowler){
    # Load IPL Bowlers
    setwd("./data")
    load("csk1.RData")
    load("dc1.RData")
    load("dd1.RData")
    load("kxip1.RData")
    load("ktk1.RData")
    load("kkr1.RData")
    load("mi1.RData")
    load("pw1.RData")
    load("rr1.RData")
    load("rcb1.RData")
    load("sh1.RData")
    load("gl1.RData")
    load("rps1.RData")
    setwd("..")
    teams_bowlers = list(csk_bowlers,dc_bowlers,dd_bowlers,kxip_bowlers,ktk_bowlers,kkr_bowlers,mi_bowlers,
                         pw_bowlers,rr_bowlers,rcb_bowlers,sh_bowlers,gl_bowlers,rps_bowlers)
    b <- NULL
    for (i in 1:length(teams_bowlers)){
        a <- which(teams_bowlers[[i]] == bowler)
        if(length(a) != 0){
            b <- c(b,i)
        }
    }
    b
}


# Get the list of the IPL team names from the indices passed
getTeams <- function(x){

    l <- NULL
    # Get the teams passed in as indexes
    for (i in seq_along(x)){

        l <- c(l, IPLTeamNames[[x[i]]]) 

    }
    l
}

# Get the team names
teamNames <- getTeams(i)

getIPLBowlerDF <- function(teamNames){
    bowlerDF <- NULL

    # Create a consolidated Data frame of batsman for all IPL teams played
    for (i in seq_along(teamNames)){
          df <- getBowlerWicketDetails(team=teamNames[i],name=IPLBowler,dir="./BattingBowlingDetails")
          bowlerDF <- rbind(bowlerDF,df) 

    }
    bowlerDF
}

14. Get the consolidated data frame for an IPL bowler

# Since an IPL bowler could have played in multiple teams we need to determine these teams and
# create a consolidated data frame for the analysis
# For example to check R Ashwin we need to do the following

IPLBowler = "R Ashwin"
#Check and get the team indices of IPL teams in which the batsman has played
i <- getTeamIndex(IPLBowler)

# Get the team names in which the IPL batsman has played
teamNames <- getTeams(i)
    # Check if file exists in the directory. This check is necessary when moving between matchType


############## Create a consolidated IPL batsman dataframe for analysis
bowlerDF <- getIPLBowlerDF(teamNames)

15. Bowler Mean Economy rate

# For e.g. to get the details of R Ashwin do
#bowlerMeanEconomyRate(bowlerDF,"R Ashwin")
bowlerMeanEconomyRate(bowlerDF,"bowlerName")

16. Bowler mean runs conceded

bowlerMeanRunsConceded(bowlerDF,"bowlerName")

17. Bowler Moving Average

bowlerMovingAverage(bowlerDF,"bowlerName")

18. Bowler cumulative average wickets

bowlerCumulativeAvgWickets(bowlerDF,"bowlerName")

19. Bowler cumulative Economy Rate (ER)

bowlerCumulativeAvgEconRate(bowlerDF,"bowlerName")

20. Bowler wicket plot

bowlerWicketPlot(bowlerDF,"bowlerName")

21. Bowler wicket against opposition

bowlerWicketsAgainstOpposition(bowlerDF,"bowlerName")

22. Bowler wicket at cricket grounds

bowlerWicketsVenue(bowlerDF,"bowlerName")

23. Predict number of deliveries to wickets

setwd("./IPLMatches")
bowlerDF1 <- getDeliveryWickets(team="IPLTeam1",dir=".",name="bowlerName",save=FALSE)
bowlerWktsPredict(bowlerDF1,"bowlerName")

Analysis of International T20 matches with yorkr templates


Introduction

In this post I create yorkr templates for International T20 matches that are available on Cricsheet. With these templates you can convert all T20 data which is in yaml format to R dataframes. Further I create data and the necessary templates for analyzing. All of these templates can be accessed from Github at yorkrT20Template. The templates are

  1. Template for conversion and setup – T20Template.Rmd
  2. Any T20 match – T20Matchtemplate.Rmd
  3. T20 matches between 2 nations – T20Matches2TeamTemplate.Rmd
  4. A T20 nations performance against all other T20 nations – T20AllMatchesAllOppnTemplate.Rmd
  5. Analysis of T20 batsmen and bowlers of all T20 nations – T20BatsmanBowlerTemplate.Rmd

Besides the templates the repository also includes the converted data for all T20 matches I downloaded from Cricsheet in Dec 2016, You can recreate the files as more matches are added to Cricsheet site. This post contains all the steps needed for T20 analysis, as more matches are played around the World and more data is added to Cricsheet. This will also be my reference in future if I decide to analyze T20 in future!

Feel free to download/clone these templates  from Github yorkrT20Template and perform your own analysis

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

There will be 5 folders at the root

  1. T20data – Match files as yaml from Cricsheet
  2. T20Matches – Yaml match files converted to dataframes
  3. T20MatchesBetween2Teams – All Matches between any 2 T20 teams
  4. allMatchesAllOpposition – A T20 countries match data against all other teams
  5. BattingBowlingDetails – Batting and bowling details of all countries
library(yorkr)
library(dplyr)

The first few steps take care of the data setup. This needs to be done before any of the analysis of T20 batsmen, bowlers, any T20 match, matches between any 2 T20 countries or analysis of a teams performance against all other countries

There will be 5 folders at the root

  1. T20data
  2. T20Matches
  3. T20MatchesBetween2Teams
  4. allMatchesAllOpposition
  5. BattingBowlingDetails

The source YAML files will be in T20Data folder

1.Create directory T20Matches

Some files may give conversions errors. You could try to debug the problem or just remove it from the T20data folder. At most 2-4 file will have conversion problems and I usally remove then from the files to be converted.

Also take a look at my Inswinger shiny app which was created after performing the same conversion on the Dec 16 data .

convertAllYaml2RDataframesT20("T20Data","T20Matches")

2.Save all matches between all combinations of T20 nations

This function will create the set of all matches between every T20 country against every other T20 country. This uses the data that was created in T20Matches, with the convertAllYaml2RDataframesT20() function.

setwd("./T20MatchesBetween2Teams")
saveAllMatchesBetweenTeams("../T20Matches")

3.Save all matches against all opposition

This will create a consolidated dataframe of all matches played by every T20 playing nation against all other nattions. This also uses the data that was created in T20Matches, with the convertAllYaml2RDataframesT20() function.

setwd("../allMatchesAllOpposition")
saveAllMatchesAllOpposition("../T20Matches")

4. Create batting and bowling details for each T20 country

These are the current T20 playing nations. You can add to this vector as more countries start playing T20. You will get to know all T20 nations by also look at the directory created above namely allMatchesAllOpposition. his also uses the data that was created in T20Matches, with the convertAllYaml2RDataframesT20() function.

setwd("../BattingBowlingDetails")
teams <-c("Australia","India","Pakistan","West Indies", 'Sri Lanka',
          "England", "Bangladesh","Netherlands","Scotland", "Afghanistan",
          "Zimbabwe","Ireland","New Zealand","South Africa","Canada",
          "Bermuda","Kenya","Hong Kong","Nepal","Oman","Papua New Guinea",
          "United Arab Emirates")

for(i in seq_along(teams)){
    print(teams[i])
    val <- paste(teams[i],"-details",sep="")
    val <- getTeamBattingDetails(teams[i],dir="../T20Matches", save=TRUE)

}

for(i in seq_along(teams)){
    print(teams[i])
    val <- paste(teams[i],"-details",sep="")
    val <- getTeamBowlingDetails(teams[i],dir="../T20Matches", save=TRUE)

}

5. Get the list of batsmen for a particular country

For e.g. if you wanted to get the batsmen of Canada you would do the following. By replacing Canada for any other country you can get the batsmen of that country. These batsmen names can then be used in the batsmen analysis

country="Canada"
teamData <- paste(country,"-BattingDetails.RData",sep="")
load(teamData)
countryDF <- battingDetails
bmen <- countryDF %>% distinct(batsman) 
bmen <- as.character(bmen$batsman)
batsmen <- sort(bmen)
batsmen

6. Get the list of bowlers for a particular country

The method below can get the list of bowler names for any T20 nation. These names can then be used in the bowler analysis below

country="Netherlands"
teamData <- paste(country,"-BowlingDetails.RData",sep="")
load(teamData)
countryDF <- bowlingDetails
bwlr <- countryDF %>% distinct(bowler) 
bwlr <- as.character(bwlr$bowler)
bowler <- sort(bwlr)
bowler

Now we are all set

A)  International T20 Match Analysis

Load any match data from the ./T20Matches folder for e.g. Afganistan-England-2016-03-23.RData

setwd("./T20Matches")
load("Afghanistan-England-2016-03-23.RData")
afg_eng<- overs
#The steps are
load("Country1-Country2-Date.Rdata")
country1_country2 <- overs

All analysis for this match can be done now

2. Scorecard

teamBattingScorecardMatch(country1_country2,"Country1")
teamBattingScorecardMatch(country1_country2,"Country2")

3.Batting Partnerships

teamBatsmenPartnershipMatch(country1_country2,"Country1","Country2")
teamBatsmenPartnershipMatch(country1_country2,"Country2","Country1")

4. Batsmen vs Bowler Plot

teamBatsmenVsBowlersMatch(country1_country2,"Country1","Country2",plot=TRUE)
teamBatsmenVsBowlersMatch(country1_country2,"Country1","Country2",plot=FALSE)

5. Team bowling scorecard

teamBowlingScorecardMatch(country1_country2,"Country1")
teamBowlingScorecardMatch(country1_country2,"Country2")

6. Team bowling Wicket kind match

teamBowlingWicketKindMatch(country1_country2,"Country1","Country2")
m <-teamBowlingWicketKindMatch(country1_country2,"Country1","Country2",plot=FALSE)
m

7. Team Bowling Wicket Runs Match

teamBowlingWicketRunsMatch(country1_country2,"Country1","Country2")
m <-teamBowlingWicketRunsMatch(country1_country2,"Country1","Country2",plot=FALSE)
m

8. Team Bowling Wicket Match

m <-teamBowlingWicketMatch(country1_country2,"Country1","Country2",plot=FALSE)
m
teamBowlingWicketMatch(country1_country2,"Country1","Country2")

9. Team Bowler vs Batsmen

teamBowlersVsBatsmenMatch(country1_country2,"Country1","Country2")
m <- teamBowlersVsBatsmenMatch(country1_country2,"Country1","Country2",plot=FALSE)
m

10. Match Worm chart

matchWormGraph(country1_country2,"Country1","Country2")

B)  International T20 Matches between 2 teams

Load match data between any 2 teams from ./T20MatchesBetween2Teams for e.g.Australia-India-allMatches

setwd("./T20MatchesBetween2Teams")
load("Australia-India-allMatches.RData")
aus_ind_matches <- matches
#Replace below with your own countries
country1<-"England"
country2 <- "South Africa"
country1VsCountry2 <- paste(country1,"-",country2,"-allMatches.RData",sep="")
load(country1VsCountry2)
country1_country2_matches <- matches

2.Batsmen partnerships

m<- teamBatsmenPartnershiOppnAllMatches(country1_country2_matches,"country1",report="summary")
m
m<- teamBatsmenPartnershiOppnAllMatches(country1_country2_matches,"country2",report="summary")
m
m<- teamBatsmenPartnershiOppnAllMatches(country1_country2_matches,"country1",report="detailed")
m
teamBatsmenPartnershipOppnAllMatchesChart(country1_country2_matches,"country1","country2")

3. Team batsmen vs bowlers

teamBatsmenVsBowlersOppnAllMatches(country1_country2_matches,"country1","country2")

4. Bowling scorecard

a <-teamBattingScorecardOppnAllMatches(country1_country2_matches,main="country1",opposition="country2")
a

5. Team bowling performance

teamBowlingPerfOppnAllMatches(country1_country2_matches,main="country1",opposition="country2")

6. Team bowler wickets

teamBowlersWicketsOppnAllMatches(country1_country2_matches,main="country1",opposition="country2")
m <-teamBowlersWicketsOppnAllMatches(country1_country2_matches,main="country1",opposition="country2",plot=FALSE)
teamBowlersWicketsOppnAllMatches(country1_country2_matches,"country1","country2",top=3)
m

7. Team bowler vs batsmen

teamBowlersVsBatsmenOppnAllMatches(country1_country2_matches,"country1","country2",top=5)

8. Team bowler wicket kind

teamBowlersWicketKindOppnAllMatches(country1_country2_matches,"country1","country2",plot=TRUE)
m <- teamBowlersWicketKindOppnAllMatches(country1_country2_matches,"country1","country2",plot=FALSE)
m[1:30,]

9. Team bowler wicket runs

teamBowlersWicketRunsOppnAllMatches(country1_country2_matches,"country1","country2")

10. Plot wins and losses

setwd("./T20Matches")
plotWinLossBetweenTeams("country1","country2")

C)  International T20 Matches for a team against all other teams

Load the data between for a T20 team against all other countries ./allMatchesAllOpposition for e.g all matches of India

load("allMatchesAllOpposition-India.RData")
india_matches <- matches
country="country1"
allMatches <- paste("allMatchesAllOposition-",country,".RData",sep="")
load(allMatches)
country1AllMatches <- matches

2. Team’s batting scorecard all Matches

m <-teamBattingScorecardAllOppnAllMatches(country1AllMatches,theTeam="country1")
m

3. Batting scorecard of opposing team

m <-teamBattingScorecardAllOppnAllMatches(matches=country1AllMatches,theTeam="country2")

4. Team batting partnerships

m <- teamBatsmenPartnershipAllOppnAllMatches(country1AllMatches,theTeam="country1")
m
m <- teamBatsmenPartnershipAllOppnAllMatches(country1AllMatches,theTeam='country1',report="detailed")
head(m,30)
m <- teamBatsmenPartnershipAllOppnAllMatches(country1AllMatches,theTeam='country1',report="summary")
m

5. Team batting partnerships plot

teamBatsmenPartnershipAllOppnAllMatchesPlot(country1AllMatches,"country1",main="country1")
teamBatsmenPartnershipAllOppnAllMatchesPlot(country1AllMatches,"country1",main="country2")

6, Team batsmen vs bowlers report

m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(country1AllMatches,"country1",rank=0)
m
m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(country1AllMatches,"country1",rank=1,dispRows=30)
m
m <-teamBatsmenVsBowlersAllOppnAllMatchesRept(matches=country1AllMatches,theTeam="country2",rank=1,dispRows=25)
m

7. Team batsmen vs bowler plot

d <- teamBatsmenVsBowlersAllOppnAllMatchesRept(country1AllMatches,"country1",rank=1,dispRows=50)
d
teamBatsmenVsBowlersAllOppnAllMatchesPlot(d)
d <- teamBatsmenVsBowlersAllOppnAllMatchesRept(country1AllMatches,"country1",rank=2,dispRows=50)
teamBatsmenVsBowlersAllOppnAllMatchesPlot(d)

8. Team bowling scorecard

teamBowlingScorecardAllOppnAllMatchesMain(matches=country1AllMatches,theTeam="country1")
teamBowlingScorecardAllOppnAllMatches(country1AllMatches,'country2')

9. Team bowler vs batsmen

teamBowlersVsBatsmenAllOppnAllMatchesMain(country1AllMatches,theTeam="country1",rank=0)
teamBowlersVsBatsmenAllOppnAllMatchesMain(country1AllMatches,theTeam="country1",rank=2)
teamBowlersVsBatsmenAllOppnAllMatchesRept(matches=country1AllMatches,theTeam="country1",rank=0)

10. Team Bowler vs bastmen

df <- teamBowlersVsBatsmenAllOppnAllMatchesRept(country1AllMatches,theTeam="country1",rank=1)
teamBowlersVsBatsmenAllOppnAllMatchesPlot(df,"country1","country1")

11. Team bowler wicket kind

teamBowlingWicketKindAllOppnAllMatches(country1AllMatches,t1="country1",t2="All")
teamBowlingWicketKindAllOppnAllMatches(country1AllMatches,t1="country1",t2="country2")

12.

teamBowlingWicketRunsAllOppnAllMatches(country1AllMatches,t1="country1",t2="All",plot=TRUE)
teamBowlingWicketRunsAllOppnAllMatches(country1AllMatches,t1="country1",t2="country2",plot=TRUE)

D) Batsman functions

Get the batsman’s details for a batsman

setwd("../BattingBowlingDetails")
kohli <- getBatsmanDetails(team="India",name="Kohli",dir=".")
batsmanDF <- getBatsmanDetails(team="country1",name="batsmanName",dir=".")

2. Runs vs deliveries

batsmanRunsVsDeliveries(batsmanDF,"batsmanName")

3. Batsman 4s & 6s

batsman46 <- select(batsmanDF,batsman,ballsPlayed,fours,sixes,runs)
p1 <- batsmanFoursSixes(batsman46,"batsmanName")

4. Batsman dismissals

batsmanDismissals(batsmanDF,"batsmanName")

5. Runs vs Strike rate

batsmanRunsVsStrikeRate(batsmanDF,"batsmanName")

6. Batsman Moving Average

batsmanMovingAverage(batsmanDF,"batsmanName")

7. Batsman cumulative average

batsmanCumulativeAverageRuns(batsmanDF,"batsmanName")

8. Batsman cumulative strike rate

batsmanCumulativeStrikeRate(batsmanDF,"batsmanName")

9. Batsman runs against oppositions

batsmanRunsAgainstOpposition(batsmanDF,"batsmanName")

10. Batsman runs vs venue

batsmanRunsVenue(batsmanDF,"batsmanName")

11. Batsman runs predict

batsmanRunsPredict(batsmanDF,"batsmanName")

12. Bowler functions

For example to get Ravicahnder Ashwin’s bowling details

setwd("../BattingBowlingDetails")
ashwin <- getBowlerWicketDetails(team="India",name="Ashwin",dir=".")
bowlerDF <- getBatsmanDetails(team="country1",name="bowlerName",dir=".")

13. Bowler Mean Economy rate

bowlerMeanEconomyRate(bowlerDF,"bowlerName")

14. Bowler mean runs conceded

bowlerMeanRunsConceded(bowlerDF,"bowlerName")

15. Bowler Moving Average

bowlerMovingAverage(bowlerDF,"bowlerName")

16. Bowler cumulative average wickets

bowlerCumulativeAvgWickets(bowlerDF,"bowlerName")

17. Bowler cumulative Economy Rate (ER)

bowlerCumulativeAvgEconRate(bowlerDF,"bowlerName")

18. Bowler wicket plot

bowlerWicketPlot(bowlerDF,"bowlerName")

19. Bowler wicket against opposition

bowlerWicketsAgainstOpposition(bowlerDF,"bowlerName")

20. Bowler wicket at cricket grounds

bowlerWicketsVenue(bowlerDF,"bowlerName")

21. Predict number of deliveries to wickets

setwd("./T20Matches")
bowlerDF1 <- getDeliveryWickets(team="country1",dir=".",name="bowlerName",save=FALSE)
bowlerWktsPredict(bowlerDF1,"bowlerName")

Inswinger: yorkr swings into International T20s


In this post I introduce ‘Inswinger’ an interactive Shiny app to analyze International T20 players, matches and teams. This app was a natural consequence to my earlier Shiny app ‘GooglyPlus’. Most of the structure for this app remained the same, I only had to work with a different dataset, so to speak.

The Googly Shiny app is based on my R package ‘yorkr’ which is now available in CRAN. The R package and hence this Shiny app is based on data from Cricsheet. Inswinger is based on the latest data dump from Cricsheet (Dec 2016) and includes all International T20 till then. There are a lot of new Internationation teams like Oman, Hong Kong, UAE, etc. In total there are 22 different International T20 teams in my Inswinger app.

The countries are a) Afghanistan b) Australia c) Bangladesh d) Bermuda e) Canada f) England g) Hong Kong h) India i) Ireland j) Kenya k) Nepal l) Netherlands m) New Zealand n) Oman o) Pakistan p) Papua New Guinea q) Scotland r) South Africa s) Sri Lanka t) United Arab Emirates u) West Indies v) Zimbabwe

My R package ‘yorkr’,  on which both these Shiny apps are based, has the ability to output either a dataframe or plot, depending on a parameter plot=TRUE or FALSE. Hence in the Inswinger Shiny app results can be displayed both as table or a plot depending on the choice of function.

Inswinger can do detailed analyses of a) Individual T20 batsman b) Individual T20 bowler c) Any T20 match d) Head to head confrontation between 2 T20 teams e) All matches of a T20 team against all other teams.

The Shiny app can be accessed at Inswinger

The code for Inswinger is available at Github. Feel free to clone/download/fork  the code from Inswinger

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

Based on the 5 detailed analysis domains there are 5 tabs
A) T20 Batsman: This tab can be used to perform analysis of all T20 batsman. If a batsman has played in more than 1 team, then the overall performance is considered. There are 10 functions for the T20 Batsman. They are shown below
– Batsman Runs vs. Deliveries
– Batsman’s Fours & Sixes
– Dismissals of batsman
– Batsman’s Runs vs Strike Rate
– Batsman’s Moving Average
– Batsman’s Cumulative Average Run
– Batsman’s Cumulative Strike Rate
– Batsman’s Runs against Opposition
– Batsman’s Runs at Venue
– Predict Runs of batsman

B) T20 Bowler: This tab can be used to analyze individual T20 bowlers. The functions handle T20 bowlers who have played in more than 1 T20 team.
– Mean Economy Rate of bowler
– Mean runs conceded by bowler
– Bowler’s Moving Average
– Bowler’s Cumulative Avg. Wickets
– Bowler’s Cumulative Avg. Economy Rate
– Bowler’s Wicket Plot
– Bowler’s Wickets against opposition
– Bowler’s Wickets at Venues
– Bowler’s wickets prediction

C) T20 match: This tab can be used for analyzing individual T20 matches. The available functions are
– Match Batting Scorecard – Table
– Batting Partnerships – Plot, Table
– Batsmen vs Bowlers – Plot, Table
– Match Bowling Scorecard   – Table
– Bowling Wicket Kind – Plot, Table
– Bowling Wicket Runs – Plot, Table
– Bowling Wicket Match – Plot, Table
– Bowler vs Batsmen – Plot, Table
– Match Worm Graph – Plot

D) Head to head: This tab can be used for analyzing head-to-head confrontations, between any 2 T20 teams for e.g. all matches between India vs Australia or West Indies vs Sri Lanka . The available functions are
-Team Batsmen Batting Partnerships All Matches – Plot, Table {Summary and Detailed}
-Team Batting Scorecard All Matches – Table
-Team Batsmen vs Bowlers all Matches – Plot, Table
-Team Wickets Opposition All Matches – Plot, Table
-Team Bowling Scorecard All Matches – Table
-Team Bowler vs Batsmen All Matches – Plot, Table
-Team Bowlers Wicket Kind All Matches – Plot, Table
-Team Bowler Wicket Runs All Matches – Plot, Table
– Win Loss All Matches – Plot

E) T20 team’s overall performance: this tab can be used analyze the overall performance of any T20 team. For this analysis all matches played by this team is considered. The available functions are
-Team Batsmen Partnerships Overall – Plot, Table {Summary and Detailed)}
-Team Batting Scorecard Overall –Table
-Team Batsmen vs Bowlers Overall – Plot, Table
-Team Bowler vs Batsmen Overall – Plot, Table
-Team Bowling Scorecard Overall – Table
-Team Bowler Wicket Kind Overall – Plot, Table

Below I include a random set of charts that are generated in each of the 5 tabs
A. IPL Batsman
a. Shakib-al-Hassan (Bangladesh) :  Runs vs Deliveries
untitled

b. Virat Kohli (India) – Cumulative Average
untitled

c.  AB Devilliers (South Africa) – Runs at venues
untitled

d. Glenn Maxwell (Australia)  – Predict runs vs deliveries faces
untitled

B. IPL Bowler
a. TG Southee (New Zealand) – Mean Economy Rate vs overs
untitled

b) DJ Bravo – Moving Average of wickets
untitled

c) AC Evans (Scotland) – Bowler Wickets Against Opposition
untitled

C.T20 Match
a. Match Score (Afghanistan vs Canada, 2012-03-18)
untitled

b)  Match batting partnerships (Plot) Hong Kong vs Oman (2015-11-21), Hong Kong
Hong Kong Partnerships
untitled

c) Match batting partnerships (Table) – Ireland vs Scotland(2012-03-18, Ireland)
Batting partnership can also be displayed as a table
untitled

d) Batsmen vs Bowlers (Plot) – India vs England (2012-12-22)
untitled

e) Match Worm Chart – Sri Lanka vs Pakistan (2015-08-01)
untitled

D.Head to head
a) Team Batsmen Partnership (Plot) – India vs Australia (all matches)
Virat Kohli has the highest total runs in partnerships against Australia
untitled

b)  Team Batsmen Partnership (Summary – Table) – Kenya vs Bangladesh
untitled

c) Team Bowling Scorecard (Table only) India vs South Africa all Matches
untitled

d) Wins- Losses New Zealand vs West Indies all Matches
untitled

C) Overall performances
a) Batting Scorecard All Matches  (Table only) – England’s overall batting performance
Eoin Morgan, Kevin Pieterson  & SJ Taylor have the best performance
untitled

b) Batsman vs Bowlers all Matches (Plot)
India’s best performing batsman (Rank=1) is Virat Kohli
untitled

c)  Batsman vs Bowlers all Matches (Table)
The plot above for Virat Kohli can also be displayed as a table. Kohli has score most runs DJ Bravo, SR Watson & Shahid Afridi
untitled

The Inswinger Shiny app can be accessed at Inswinger. Give it a swing!

The code for Inswinger is available at Github. Feel free to clone/download/fork  the code from Inswinger

Also see my other Shiny apps
1.GooglyPlus
2.What would Shakespeare say?
3.Sixer
4.Revisiting crimes against women in India

You may also like
1. Neural Networks: The mechanics of backpropagation
A primer on Qubits, Quantum gates and Quantum Operation
2. Re-working the Lucy Richardson algorithm in OpenCV
3.Design Principles of Scalable, Distributed Systems
4.Spicing up a IBM Bluemix cloud app with MongoDB and NodeExpress
5.Programming languages in layman’s language
7.Re-introducing cricketr! : An R package to analyze performances of cricketers

To see all posts take at a look at Index of Posts

Neural Networks: On Perceptrons and Sigmoid Neurons


Neural Networks had their beginnings in 1943 when Warren McCulloch, a neurophysiologist, and a young mathematician, Walter Pitts, wrote a paper on how neurons might work.  Much later in 1958, Frank Rosenblatt, a neuro-biologist proposed the Perceptron. The Perceptron is a computer model or computerized machine which is devised to represent or simulate the ability of the brain to recognize and discriminate. In machine learning, the perceptron is an algorithm for supervised learning of binary classifiers

Initially it was believed that  Perceptrons were capable of many things including “the ability to walk, talk, see, write, reproduce itself and be conscious of its existence.”

However, a subsequent paper by Marvin Minky and Seymour Papert of MIT, titled “Perceptrons” proved that the Perceptron was truly limited in its functionality. Specifically they showed that the Perceptron was incapable of producing XOR functionality. The Perceptron is only capable of classification where the data points are linearly separable.

This post implements the simple learning algorithm of the ‘Linear Perceptron’ and the ‘Sigmoid Perceptron’.  The implementation has been done in Octave. This implementation is based on “Neural networks for Machine Learning” course by Prof Geoffrey Hinton at Coursera

Perceptron learning procedure
z = ∑wixi  + b
where wi is the ith weight and xi is the ith  feature

For every training case compute the activation output zi

  • If the output classifies correctly, leave the weights alone
  • If the output classifies a ‘0’ as a ‘1’, then subtract the the feature from the weight
  • If the output classifies a ‘0’ as a ‘1’, then add the feature to the weight

This simple neural network is represented below
perceptron

Sigmoid neuron learning procedure
zi = sigmoid(∑wixi  + b)
where sigmoid is
sigmoid(z) = 1/1+e^{-z}

Hence
z_{i} = 1/1+e^{-(\sum w_{i}x_{i}+b)}
For every training case compute the activation output zi

  • If the output classifies correctly, leave the weights alone
  • If the output incorrectly classifies a ‘0’ as a ‘1’ i.e. z_{i} >sigmoid(0), then subtract the feature from the weight
  • If the output incorrectly classifies a ‘1’ as ‘0’ i.e., i.e z_{i} < sigmoid(0), then add the feature to the weight
  • Iterate till errors <= 1

This is shown below
sigmoid_neuron

I have implemented the learning algorithm of the Perceptron and Sigmoid Neuron in Octave. The code is available at Github at Perceptron.

  1. Perceptron execution

I performed the tests on 2 different datasets

Data 1
untitled

Data 2
untitled

2. Sigmoid Perceptron execution
Data 1 & Data 2

It can be seen that the Perceptron does work for simple linearly separable data. I will be implementing other more advanced Neural Networks in the months to come.

Watch this space!

GooglyPlus: yorkr analyzes IPL players, teams, matches with plots and tables


In this post I introduce my new Shiny app,“GooglyPlus”, which is a  more evolved version of my earlier Shiny app “Googly”. My R package ‘yorkr’,  on which both these Shiny apps are based, has the ability to output either a dataframe or plot, depending on a parameter plot=TRUE or FALSE. My initial version of the app only included plots, and did not exercise the yorkr package fully. Moreover, I am certain, there may be a set of cricket aficionados who would prefer, numbers to charts. Hence I have created this enhanced version of the Googly app and appropriately renamed it as GooglyPlus. GooglyPlus is based on the yorkr package which uses data from Cricsheet. The app is based on IPL data from  all IPL matches from 2008 up to 2016. Feel free to clone/fork or download the code from Github at GooglyPlus.

Click  GooglyPlus to access the Shiny app!

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

The changes for GooglyPlus over the earlier Googly app is only in the following 3 tab panels

  • IPL match
  • Head to head
  • Overall Performance

The analysis of IPL batsman and IPL bowler tabs are unchanged. These charts are as they were before.

The changes are only in  tabs i) IPL match ii) Head to head and  iii) Overall Performance. New functionality has been added and existing functions now have the dual option of either displaying a plot or a table.

The changes are

A) IPL Match
The following additions/enhancements have been done

-Match Batting Scorecard – Table
-Batting Partnerships – Plot, Table (New)
-Batsmen vs Bowlers – Plot, Table(New)
-Match Bowling Scorecard   – Table (New)
-Bowling Wicket Kind – Plot, Table (New)
-Bowling Wicket Runs – Plot, Table (New)
-Bowling Wicket Match – Plot, Table (New)
-Bowler vs Batsmen – Plot, Table (New)
-Match Worm Graph – Plot

B) Head to head
The following functions have been added/enhanced

-Team Batsmen Batting Partnerships All Matches – Plot, Table {Summary (New) and Detailed (New)}
-Team Batting Scorecard All Matches – Table (New)
-Team Batsmen vs Bowlers all Matches – Plot, Table (New)
-Team Wickets Opposition All Matches – Plot, Table (New)
-Team Bowling Scorecard All Matches – Table (New)
-Team Bowler vs Batsmen All Matches – Plot, Table (New)
-Team Bowlers Wicket Kind All Matches – Plot, Table (New)
-Team Bowler Wicket Runs All Matches – Plot, Table (New)
-Win Loss All Matches – Plot

C) Overall Performance
The following additions/enhancements have been done in this tab

-Team Batsmen Partnerships Overall – Plot, Table {Summary (New) and Detailed (New)}
-Team Batting Scorecard Overall –Table (New)
-Team Batsmen vs Bowlers Overall – Plot, Table (New)
-Team Bowler vs Batsmen Overall – Plot, Table (New)
-Team Bowling Scorecard Overall – Table (New)
-Team Bowler Wicket Kind Overall – Plot, Table (New)

Included below are some random charts and tables. Feel free to explore the Shiny app further

1) IPL Match
a) Match Batting Scorecard (Table only)
This is the batting score card for the Chennai Super Kings & Deccan Chargers 2011-05-11

untitled

b)  Match batting partnerships (Plot)
Delhi Daredevils vs Kings XI Punjab – 2011-04-23

untitled

c) Match batting partnerships (Table)
The same batting partnership  Delhi Daredevils vs Kings XI Punjab – 2011-04-23 as a table

untitled

d) Batsmen vs Bowlers (Plot)
Kolkata Knight Riders vs Mumbai Indians 2010-04-19

Untitled.png

e)  Match Bowling Scorecard (Table only)
untitled

B) Head to head

a) Team Batsmen Partnership (Plot)
Deccan Chargers vs Kolkata Knight Riders all matches

untitled

b)  Team Batsmen Partnership (Summary – Table)
In the following tables it can be seen that MS Dhoni has performed better that SK Raina  CSK against DD matches, whereas SK Raina performs better than Dhoni in CSK vs  KKR matches

i) Chennai Super Kings vs Delhi Daredevils (Summary – Table)

untitled

ii) Chennai Super Kings vs Kolkata Knight Riders (Summary – Table)
untitled

iii) Rising Pune Supergiants vs Gujarat Lions (Detailed – Table)
This table provides the detailed partnership for RPS vs GL all matches

untitled

c) Team Bowling Scorecard (Table only)
This table gives the bowling scorecard of Pune Warriors vs Deccan Chargers in all matches

untitled

C) Overall performances
a) Batting Scorecard All Matches  (Table only)

This is the batting scorecard of Royal Challengers Bangalore. The top 3 batsmen are V Kohli, C Gayle and AB Devilliers in that order

untitled

b) Batsman vs Bowlers all Matches (Plot)
This gives the performance of Mumbai Indian’s batsman of Rank=1, which is Rohit Sharma, against bowlers of all other teams

untitled

c)  Batsman vs Bowlers all Matches (Table)
The above plot as a table. It can be seen that Rohit Sharma has scored maximum runs against M Morkel, then Shakib Al Hasan and then UT Yadav.

untitled

d) Bowling scorecard (Table only)
The table below gives the bowling scorecard of CSK. R Ashwin leads with a tally of 98 wickets followed by DJ Bravo who has 88 wickets and then JA Morkel who has 83 wickets in all matches against all teams

Untitled.png

This is just a random selection of functions. Do play around with the app and checkout how the different IPL batsmen, bowlers and teams stack against each other. Do read my earlier post Googly: An interactive app for analyzing IPL players, matches and teams using R package yorkr  for more details about the app and other functions available.

Click GooglyPlus to access the Shiny app!

You can clone/fork/download the code from Github at GooglyPlus

Hope you have fun playing around with the Shiny app!

Note: In the tabs, for some of the functions, not all controls  are required. It is possible to enable the controls selectively but this has not been done in this current version. I may make the changes some time in the future.

Take a look at my other Shiny apps
a.Revisiting crimes against women in India
b. Natural language processing: What would Shakespeare say?

Check out some of my other posts
1. Analyzing World Bank data with WDI, googleVis Motion Charts
2. Video presentation on Machine Learning, Data Science, NLP and Big Data – Part 1
3. Singularity
4. Design principles of scalable, distributed systems
5. Simulating an Edge shape in Android
6. Dabbling with Wiener filter in OpenCV

To see all posts click Index of Posts

yorkr ranks IPL Players post 2016 season


Here is a short post which ranks IPL batsmen and bowlers post the 2016 IPL season. These are based on match data from Cricsheet. I had already ranked IPL players in my post yorkr ranks IPL batsmen and bowlers, but that was mid IPL 2016 season. This post will be final ranking post 2016 season

This post has also been published in RPubs RankIPLPlayers2016. You can download this as a pdf file at RankIPLPlayers2016.pdf.

You can take a look at the code at rankIPLPlayers2016

Check out my 2 books on cricket, a) Cricket analytics with cricketr b) Beaten by sheer pace – Cricket analytics with yorkr, now available in both paperback & kindle versions on Amazon!!! Pick up your copies today!

Checkout my interactive Shiny apps GooglyPlus (plots & tables) and Googly (only plots) which can be used to analyze IPL players, teams and matches.

rm(list=ls())
library(yorkr)
library(dplyr)
source('C:/software/cricket-package/cricsheet/ipl2016/final/R/rankIPLBatsmen.R', encoding = 'UTF-8')
source('C:/software/cricket-package/cricsheet/ipl2016/final/R/rankIPLBowlers.R', encoding = 'UTF-8')

Rank IPL batsmen post 2016

Chris Gayle, Shaun Marsh & David Warner are top 3 IPL batsmen. Gayle towers over everybody, with an 38.28 Mean Runs, and a Mean Strike Rate of 138.85. Virat Kohli comes in 4th, with 34.52 as his Average Runs per innings, and a Mean Strike Rate of 117.51

iplBatsmanRank <- rankIPLBatsmen()
as.data.frame(iplBatsmanRank[1:30,])
##             batsman matches meanRuns    meanSR
## 1          CH Gayle      92 38.28261 138.85120
## 2          SE Marsh      60 36.40000 118.97783
## 3         DA Warner     104 34.51923 124.88798
## 4           V Kohli     136 31.77941 117.51000
## 5         AM Rahane      89 31.46067 104.62989
## 6    AB de Villiers     109 29.93578 136.48945
## 7      SR Tendulkar      78 29.62821 108.58962
## 8         G Gambhir     133 28.94737 109.61263
## 9         RG Sharma     140 28.68571 117.79057
## 10         SK Raina     143 28.41259 121.55713
## 11        SR Watson      90 28.21111 125.80122
## 12         S Dhawan     110 28.09091 111.97282
## 13         R Dravid      79 27.87342 109.14544
## 14         DR Smith      76 27.55263 120.22329
## 15        JP Duminy      70 27.28571 122.99243
## 16      BB McCullum      94 26.86170 118.55606
## 17        JH Kallis      97 26.83505  95.47866
## 18         V Sehwag     105 26.26667 137.11562
## 19       RV Uthappa     132 26.18182 123.16326
## 20     AC Gilchrist      81 25.77778 122.69074
## 21          M Vijay      99 25.69697 106.02010
## 22    KC Sangakkara      70 25.67143 112.97529
## 23         MS Dhoni     131 25.14504 131.62206
## 24        DA Miller      60 24.76667 133.80983
## 25        AT Rayudu      99 23.35354 121.59313
## 26 DPMD Jayawardene      80 23.05000 114.54712
## 27     Yuvraj Singh     103 22.46602 118.15000
## 28        DJ Hussey      63 22.26984        NA
## 29        YK Pathan     121 22.25620 132.58793
## 30      S Badrinath      66 22.22727 114.97061

Rank IPL bowlers

The top 3 IPL T20 bowlers are SL Malinga, DJ Bravo and SP Narine

Don’t get hung up on the decimals in the average wickets for the bowlers. All it implies is that if 2 bowlers have average wickets of 1.0 and 1.5, it implies that in 2 matches the 1st bowler will take 2 wickets and the 2nd bowler will take 3 wickets.

setwd("C:/software/cricket-package/cricsheet/ipl2016/details")
iplBowlersRank <- rankIPLBowlers()
as.data.frame(iplBowlersRank[1:30,])
##             bowler matches meanWickets   meanER
## 1       SL Malinga      96    1.645833 6.545208
## 2         DJ Bravo      58    1.517241 7.929310
## 3        SP Narine      65    1.492308 6.155077
## 4          B Kumar      45    1.422222 7.355556
## 5        YS Chahal      41    1.414634 8.057073
## 6         M Morkel      37    1.405405 7.626216
## 7        IK Pathan      40    1.400000 7.579250
## 8         RP Singh      42    1.357143 7.966429
## 9         MM Patel      31    1.354839 7.282581
## 10   R Vinay Kumar      63    1.317460 8.342540
## 11  Sandeep Sharma      38    1.315789 7.697368
## 12       MM Sharma      46    1.304348 7.740652
## 13         P Awana      33    1.303030 8.325758
## 14        MM Patel      30    1.300000 7.569667
## 15          Z Khan      41    1.292683 7.735854
## 16         PP Ojha      53    1.245283 7.268679
## 17     JP Faulkner      40    1.225000 8.502250
## 18 Shakib Al Hasan      41    1.170732 7.103659
## 19     DS Kulkarni      32    1.156250 8.372188
## 20        UT Yadav      46    1.152174 8.394783
## 21        A Kumble      41    1.146341 6.567073
## 22       JA Morkel      73    1.136986 8.131370
## 23        SK Warne      53    1.132075 7.277170
## 24        A Mishra      55    1.127273 7.319455
## 25        UT Yadav      33    1.090909 8.853636
## 26        L Balaji      34    1.088235 7.186176
## 27       PP Chawla      35    1.085714 8.162000
## 28        R Ashwin      92    1.065217 6.812391
## 29  M Muralitharan      39    1.051282 6.470256
## 30 Harbhajan Singh     120    1.050000 7.134833