R studio how to…

GENERAL COMMANDS ................................................................................................................................................ 3

ATTACH A DATASET ............................................................................................................................................................... 3
SAVE A DATASET ................................................................................................................................................................... 3
CREATE A NEW DATASET FROM AN EXISTING DATASET ................................................................................................................. 3
DELETE VARIABLES FROM A DATASET ........................................................................................................................................ 3
EXCLUDE NA FROM A DATASET ............................................................................................................................................... 3
COMBINE DATAFRAMES ......................................................................................................................................................... 3
RENAME A DATASET .............................................................................................................................................................. 3
ESTIMATE THE LOGARITHM OF A SERIES .................................................................................................................................... 3
RECODE A CATEGORICAL VARIABLE ........................................................................................................................................... 3

SUMMARY STATISTICS COMMANDS ............................................................................................................................ 4

PRODUCE SUMMARY STATISTICS .............................................................................................................................................. 4
ESTIMATE THE MEAN ............................................................................................................................................................. 4
ESTIMATE THE VARIANCE........................................................................................................................................................ 4
ESTIMATE THE STANDARD DEVIATION ....................................................................................................................................... 4

THE NORMAL AND THE STUDENT- T DISTRIBUTION ..................................................................................................... 5

ESTIMATE THE PROBABILITY DENSITY FUNCTION FOR THE NORMAL DISTRIBUTION ............................................................................. 5
ESTIMATE THE INTEGRAL FROM −∞ TO Q OF THE PDF OF THE NORMAL DISTRIBUTION WHERE Q IS A Z-SCORE ........................................ 5
ESTIMATE THE Z-SCORE OF THE PTH QUANTILE OF THE NORMAL DISTRIBUTION ................................................................................ 5
ESTIMATE A VECTOR OF NORMALLY DISTRIBUTED RANDOM NUMBERS ............................................................................................ 5
ESTIMATE THE DENSITY FUNCTION FOR THE T DISTRIBUTION WITH DF DEGREES OF FREEDOM.............................................................. 5
ESTIMATE THE DISTRIBUTION FUNCTION FOR THE T DISTRIBUTION WITH DF DEGREES OF FREEDOM ...................................................... 5
ESTIMATE THE QUANTILE FUNCTION FOR THE T DISTRIBUTION WITH DF DEGREES OF FREEDOM ............................................................ 5
ESTIMATE RANDOM NUMBERS WITH T DISTRIBUTION .................................................................................................................. 5

CONFIDENCE INTERVALS.............................................................................................................................................. 6
ESTIMATE A 90% CONFIDENCE INTERVAL ................................................................................................................................. 6
ESTIMATE A 95% CONFIDENCE INTERVAL ................................................................................................................................. 6
ESTIMATE A 99% CONFIDENCE INTERVAL ................................................................................................................................. 6
ESTIMATE A P-VALUE T-DISTRIBUTION ...................................................................................................................................... 6

PLOTS COMMANDS ..................................................................................................................................................... 7

PRODUCE A HISTOGRAM ........................................................................................................................................................ 7
ADJUST THE X- AND Y-AXIS OF A HISTOGRAM ............................................................................................................................. 7
ADJUST THE NUMBER OF BINS OF A HISTOGRAM ......................................................................................................................... 7
PRODUCE A SCATTER PLOT ...................................................................................................................................................... 7
ADD A MAIN TITLE IN A SCATTER PLOT ...................................................................................................................................... 7
ADD THE REGRESSION LINE IN A SCATTER PLOT ........................................................................................................................... 7
PLOT TWO GRAPHS SIDE BY SIDE .............................................................................................................................................. 7
PLOT TWO GRAPHS ON THE SAME AXIS ..................................................................................................................................... 7
PLOT A GRAPH AND ADD A LEGEND .......................................................................................................................................... 7

CORRELATION MATRIX COMMANDS ........................................................................................................................... 8

ESTIMATE THE CORRELATION MATRIX ....................................................................................................................................... 8

REGRESSION COMMANDS ........................................................................................................................................... 9

ESTIMATE AN OLS MODEL...................................................................................................................................................... 9
STORE THE RESIDUALS OF AN OLS MODEL ................................................................................................................................. 9
ESTIMATE ROBUST STANDARD ERRORS ..................................................................................................................................... 9
ACCESS IF TWO VARIABLES ARE JOINTLY SIGNIFICANT ................................................................................................................... 9

Dr. Maria Giamouzi 1

RESIDUAL DIAGNOSTIC COMMANDS ......................................................................................................................... 10
ESTIMATE THE WHITE TEST - TESTING FOR HETEROSKEDASTICITY ................................................................................................. 10
ESTIMATE WHITE TEST - TESTING FOR HETEROSKEDASTICITY ....................................................................................................... 10
ESTIMATE BREUSCH AND PAGAN TEST - TESTING FOR HETEROSKEDASTICITY .................................................................................. 10
PLOT THE REGRESSION LINE AND THE DATA .............................................................................................................................. 10
ESTIMATE THE RESET TEST .................................................................................................................................................. 11
ESTIMATE JARQUE-BERA TEST TO ACCESS THE NORMALITY OF THE RESIDUALS ................................................................................ 11
ESTIMATE THE AIC VALUES ................................................................................................................................................... 11
ESTIMATE THE BIC VALUES ................................................................................................................................................... 11

CROSS VALIDATION COMMANDS .............................................................................................................................. 12

CONSTRUCT THE IN-SAMPLE AND OUT-OF-SAMPLE PERIOD ......................................................................................................... 12
ESTIMATE THE OLS MODEL USING THE IN-SAMPLE AND OUT-OF-SAMPLE PERIOD ........................................................................... 12

Dr. Maria Giamouzi 2

General Commands

attach a dataset
attach(name of the dataset)
example – attach(data)

save a dataset
save(name of the new dataset,file="name of the file.Rda")
example - save(combined,file="inc_literacy_2016.Rda")

create a new dataset from an existing dataset

NameOfTheNewDataset <- ExistingDatasetName[,c(“NameOfVariable#1”, “NameOfVariable#2”,…)]
example - temp <- income[,c("Country.Name","X2016")]

delete variables from a dataset

NameOfTheNewDataset <- subset(NameOfExistingDataset = -c(“NameOfVariable#1ToBeEliminated”,
“NameOfVariable#2 ToBeEliminated”,…)
example – hprices2 <- subset(hprices1=-c(`prices`))

exclude NA from a dataset

NameOfDataset_NA <- na.exclude(NameOfDataset)

combine dataframes
NewDatasetName <- cbind(NameOfDataframe1, NameOfDataframe1,..)
example - combined <- cbind(temp,literacy[,"X2016"])

rename a dataset
NewName <- OldName

estimate the logarithm of a series

New_variable_name <- log(Name_of_the_Series)

recode a categorical variable

library(car)
accepted <- recode(Variable_Name, " 'yes' = 0; 'no' = 1; ", as.factor = FALSE)

Dr. Maria Giamouzi 3

Summary Statistics Commands

produce summary statistics

summary (Name_of_dataset)

estimate the mean

mean(Name_of_variable, na.rm=TRUE)

estimate the variance

var(Name_of_variable, na.rm = TRUE)

estimate the standard deviation

sqrt(var(Name_of_variable, na.rm = TRUE))
or
sd(Name_of_variable, na.rm = TRUE)

Dr. Maria Giamouzi 4

The Normal and the Student- t Distribution

estimate the probability density function for the normal distribution

dnorm(state the value)

estimate the integral from −∞ to q of the pdf of the normal distribution where q
is a Z-score
pnorm(value)

estimate the Z-score of the pth quantile of the normal distribution

qnorm(value)

estimate a vector of normally distributed random numbers

rnorm(number of random numbers to be generated,mean,sd)
example - rnorm(10,mean=5,sd = 2)

estimate the density function for the t distribution with df degrees of freedom
dt(value, df)

estimate the distribution function for the t distribution with df degrees of

freedom
pt(value, df)

estimate the quantile function for the t distribution with df degrees of freedom
qt(value, df)

estimate random numbers with t distribution

rt(10,20)

Dr. Maria Giamouzi 5

Confidence Intervals

estimate a 90% Confidence Interval

ci = 0.90
dfs = state the degrees of freedom
tval <- qt((ci +0.5*(1 - ci)), dfs)

upper_bound <- coeeficient + tval*Standard_Error

lower_bound <- coeeficient - tval*Standard_Error

estimate a 95% Confidence Interval

ci = 0.95
dfs = 38
tval <- qt((ci +0.5*(1 - ci)), dfs)
upper_bound <- coeeficient + tval*Standard_Error
lower_bound <- coeeficient - tval*Standard_Error

estimate a 99% Confidence Interval

ci = 0.99
dfs = 38
tval <- qt((ci +0.5*(1 - ci)), dfs)
upper_bound <- coeeficient + tval*Standard_Error
lower_bound <- coeeficient - tval*Standard_Error

estimate a p-value t-distribution

pvalue <- pt(t_statistic_value, degrees_of_freedom, lower.tail = TRUE)
pvalue <- pt(t_statistic_value, degrees_of_freedom, lower.tail = FALSE)

Dr. Maria Giamouzi 6

Plots commands

produce a histogram
hist(name_of_variable)

adjust the x- and y-axis of a histogram

hist(name_of_variable, xlim=c(lower value,max value), ylim=c(lower value,max value))

adjust the number of bins of a histogram

hist(name_of_variable, breaks=number_of_bins)

produce a scatter plot

plot(name_of_variable_x, name_of_variable_y)

add a main title in a scatter plot

plot(name_of_variable_x, name_of_variable_y, main="TITLE")

add the regression line in a scatter plot

abline(name_of_variable_y~name_of_variable_x)

plot two graphs side by side

par(mfrow=c(1,2))
hist(variable name 1) hist(variable name 2)

plot two graphs on the same axis

par(mfrow=c(1,1)) # set the plot to a single axes
p1 <- hist(variable name 1)
p2 <- hist(variable name 2)
plot(p1, col=“green”, xlim =c(lower bound,upper bound), main = "TITLE")
#the col option specifies the colouring of the bars
plot(p2, col= “red”, xlim=c(lower bound,upper bound), add=T)
#the add=T option adds the second distribution onto the existing plot

plot a graph and add a legend

png(filename="Name_of_Graph.png")
#this line, prior to the plotting commands, leads to the plot being saved in #the current directory, with the specified
filename and format
plot(Name_of_variable, type = "l", col = "green", main = "TITLE")
lines(Name_of_variable, type = "l", col = "black")
legend("topright", legend=c("Name of Variable 1", " Name of Variable 2"), col=c("green", "black"), lwd = 2)
dev.off() # this line `finishes' the command to save the plot to the file named above

Dr. Maria Giamouzi 7

Correlation Matrix Commands

estimate the correlation matrix

install.packages (“corrplot”)
Correlation_Matrix <- cor(na.omit(Name_of_dataset))

library(corrplot)
corrplot(Correlation_Matrix,method="number",type="lower")
or
corrplot(Correlation_Matrix,method="number",type="upper")
or
corrplot(Correlation_Matrix,method="circle")
or
corrplot(Correlation_Matrix,method="pie")
or
corrplot(Correlation_Matrix,method=="color")

Here you can find additional commands on how to modify your correlation matrix.

Dr. Maria Giamouzi 8

Regression Commands

estimate an OLS model

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

store the residuals of an OLS model

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

resids <- resid(Name_the_model)

estimate robust standard errors

cov1 <- hccm(Name_the_model)
robustse <- sqrt(diag(cov1))
standard_se <- sqrt(diag(vcov(eq1)))
t_values_with_RSE <- Name_the_model$coefficients/robustse
pvalues <- pt(Name_the_model$coefficients/robustse, df, lower.tail = FALSE)

access if two variables are jointly significant

linearHypothesis(Name_the_model, c("Variable_name_1=0", " Variable_name_1=0"))

or
H0 <- c("Variable_name_1 = 0", " Variable_name_2= 0")
#we have specified each parameter = 0 to separately, but the function runs the test of both jointly, so it is a
joint test
linearHypothesis(Name_the_model, H0)

Dr. Maria Giamouzi 9

Residual Diagnostic Commands

estimate the White test - testing for heteroskedasticity

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

resids <- resid(Name_the_model)

sqres < resids^2

white1 <- lm(sqres ~ independent variable_1 + independent variable_2,… independent variable_1^2,…,

independent variable_1* independent variable_2,…)
summary(white1)

estimate White test - testing for heteroskedasticity

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

yhat <- Name_the_model$fitted.values

white2 <- lm(sqres ~ yhat + I(yhat^2))
summary(white2)

estimate Breusch and Pagan Test - testing for heteroskedasticity

library(wooldridge)
library(car)
name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)
ncvTest(Name_the_model)

H0: the residuals are homoscedastic

H1: the residuals are heteroscedastic

If the p-value is less than the significance level we have to reject the null hypothesis, therefore that implies
that the residuals are heteroscedastic.

If the p-value is greater than the significance level we have to accept the null hypothesis, therefore that
implies that the residuals are homoscedastic.

plot the regression line and the data

plot(independent_variable, dependent_variable, main="TITLE")
abline(dependent_variable ~ independent_variable)

Dr. Maria Giamouzi 10

 estimate the RESET test
name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

resids <- Name_the_model $residuals

yhats <- Name_the_model $fitted.values
reseteq <- lm(resids ~ I(yhats^2))
summary(reseteq)
reseteq2 <- lm(resids ~ I(yhats^2) + I(yhats^3))
summary(reseteq2)

estimate Jarque-Bera test to access the normality of the residuals

install.packages(“xts”)
install.packages (“tseries”)

library(tseries)
name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)

resids <- Name_the_model $residuals

jarque.bera.test(resids)

H0: The residuals are normally distributed

H1: The residuals are not normally distributed

If the p-value is less than the significance level we have to reject the null hypothesis, therefore that implies
that the residuals are not normally distributed.
If the p-value is greater than the significance level we have to accept the null hypothesis, therefore that
implies that the residuals are normally distributed.

estimate the AIC values

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)
library(stats)
aic <- AIC(Name_the_model)

estimate the BIC values

name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…)
summary (Name_the_model)
library(nlme)
bic <- BIC(Name_the_model)

Dr. Maria Giamouzi 11

Cross Validation Commands

construct the in-sample and out-of-sample period

In_Sample_Period <- Dataset_Name[1:200,]
Out_of_Sample_Period <- Dataset_Name [-(1:200),]

200 is just a random number that needs to be adjusted depending on the data under investigation.

estimate the OLS model using the in-sample and out-of-sample period
name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…, data=
In_Sample_Period)
summary (Name_the_model)

# Then use this model to predict the “out-of-sample” values

ypred2 <- predict.lm(name_the_model, newdata = Out_of_Sample_Period)
errors2 <- In_Sample_Period $Dependent_Variable - ypred2
plot(errors2)

or
name_the_model < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…, data=
In_Sample_Period)
summary (Name_the_model)
errors <- resids (name_the_model)

name_the_model_2 < - lm(dependent_variable ~ independent variable_1 + independent variable_2,…, data=

Out_of_Sample_Period)
summary (Name_the_model_2)
resids <- resids (name_the_model_2)

sde1 <- sqrt(var(errors))

sde2 <- sqrt(var(resids))

Dr. Maria Giamouzi 12