Statistical

PART I: Analysis
Introduction & using
Data Exploration
5th December 2020
Nur Saadah Binti Abd Majid 10am-1.00pm
10am-11.30am Via Microsoft Teams
 Course Outline
Importing Learn how to import data with different type such
Dataset
as csv, text, stata and excel.

Understanding
the dataset Use the basic built in function to get
knowledge on the data

Data Explore the data using the plots that

Exploration
basically required for EDA process.

Exporting Save and export the

Dataset
updated data.

Exercise Measure the

understanding.
Setting Working Directory

setwd("<working path>")

OR
File Change dir… <Find your folder>
 Importing the dataset
Dataset data<-
in read.table("mydata.txt
text ",sep=";")
file

data<- Dataset
in
read.csv("mydata.csv",
header=T, sep=",") csv
file

Dataset
library(foreign)
in
data <-
Stata read.dta("mydata.dta")
file
 Built-in dataset

Example of datasets

data(mtcars)
help(mtcars)
data(iris)
help(iris)

List of all built-in dataset

data()
Generating normally
distributed data
#1. density function
dnorm(0.5,mean = 0, sd = 1, log =
FALSE)

#2. distribution function

pnorm(0.5, mean = 0, sd = 1, lower.tail
= TRUE, log.p = FALSE)

#3. quantile function

qnorm(p, mean = 0, sd = 1, lower.tail =
TRUE, log.p = FALSE)

#4. Generate a vector of normally

distributed random numbers
set.seed(1005)
dat<-rnorm(100, mean = 0, sd = 1)
dat<-round(dat,2)
 Understanding the
structure of the data

USING VECTORS USING DATASETS

head(data)
a<-rep(2,15) tail(data)
b<-rep(1:3,5) nrow(data)
ncol(data)
c<-rep(c(3,5,10:15),2)
dim(data)
d<-c(1:15) colnames(data)
try1<-cbind(a,b) rownames(data)
str(data)
try2<-rbind(a,b)
summary(data)
try3<-c(a,c)
 Checking the missing
values/duplicate Check the
duplicated
rows

CHECKING THE
DUPLICATES
class(c)
class(data)
class(try1)
#to make try1 become data.frame
try1<-as.data.frame(try1)
class(try1)

#Find the missing value

library(dplyr)
duplicated(try1)
distinct(try1,try1$b)
Gives the
unique(try1) distinct
value for a
variable
#Identifying rows that duplicate
dup.rownum<-which(duplicated(try1)=="TRUE")
try1[dup.rownum,] Removes the
duplicated
rows
automatically
 Checking the missing
values/duplicate
CHECKING THE
MISSING VALUES

Create data
with Missing
#Find the MISSING VALUES Values
try1[c(3,6),1]<-NA
try1[c(10,15),2]<-NA
try1
complete.cases(try1)
mis.rownum<-which(complete.cases(try1)=="FALSE“)
try1[mis.rownum,]

Check the
rows without
the missing
values
Sorting and Ordering the
data Sorting the
vector
automatically

Give the
number of
order for
each rows

sort(b)
sort(b,decreasing=T)
order(b)
order(b,decreasing=T)
try1<-cbind(a,b)
Extract the
try1<-as.data.frame(try1) data by order

try1[order(b),]
 If else
Dataset in a
vector type

avr<-mean(data[,2])
if (data[5,2]>avr) {
print("Larger than mean")
} The value
else { assigned is
larger than
print("Smaller than mean") the mean of
} the vector No

"Smaller
Yes than mean"

"Larger than
mean"
 Looping
i-th observation
of dataset in a
vector type

The ith

"Smaller
No value is Yes "Larger
larger
than the
than
than mean" mean"
mean of
the vector
for(i in 1:nrow(data)){
if (data[i,2]>avr) {
print("Larger than mean")
} else { The i value is
print("Smaller than mean") increasing by 1
} and create new
} value of i

The i value
is larger
No
than the
max value

Yes

Terminate
 Data Exploration &
Plottings

SCATTER PLOT

help(mtcars)
summary(mtcars)
dim(mtcars)

plot(mtcars[,1],mtcars[,3],ylim
=c(min(mtcars[,3]),max(mtcars[, The Displacement against Miles/US gallon
3])),xlim=c(min(mtcars[,1]),max
(mtcars[,1])),ylab="Displacemen
t (cu.in.)",xlab="Miles/US
400

gallon", main="The Displacement

against Miles/US gallon")
Displacement (cu.in.)

300

abline(v=mean(mtcars[,1]),
lty=2, col="Red")
200

abline(h=mean(mtcars[,3]),
lty=5, col="Blue")
100

points(mtcars[1:5,1],mtcars[1:5
,3],col="Purple") 10 15 20 25 30

Miles/US gallon

Save as Save as
metafile bitmap
 Data Exploration &
Plottings

LINE PLOT
plot(sunspots[1:200],main="Monthly averaged sunspots
from 1749–1983",xaxt="n",xlabs="Month order",
ylab="Sunspots number")

lines(sunspots[1:200],lty=1,col="Purple")

lines(sunspots[1:200]+20,lty=1,col="Blue")

axis(1,1:200,1:200,las=2)

legend("topright", legend=c("Sunspot", "Sunspot+20"),

col=c("purple", "blue"),lty=c(1,1),cex=0.8)

TIME SERIES PLOT

ts.plot(sunspots[1:200],main="Monthly
averaged sunspots from 1749–1983",
ylab="Sunspots number",col="blue")
Data Exploration &
Plottings

HISTOGRAM

hist(mtcars[,1],xlab
= "Weight",col =
"pink",border =
"red", main="Miles/US
gallon", breaks =
10)
Data Exploration &
Plottings

BOX PLOT

boxplot(mtcars[,1],main="Miles/U
S gallon", horizontal=T)
boxplot(mtcars[,3],main=
"Displacement (cu.in.)")

par(mfrow=c(2,1))
boxplot(mtcars[,1],main=
"Miles/US gallon", horizontal=T)

boxplot(mtcars[,3],main=
"Displacement (cu.in.)",
horizontal=T)

par(mfrow=c(1,2))
boxplot(mtcars[,1],main=
"Miles/US gallon" )

boxplot(mtcars[,3],main=
"Displacement (cu.in.)")
Data Exploration &
Plottings

QQPLOT

qqnorm(mtcars[,1],
pch = 1, frame =
FALSE)

qqline(mtcars[,1],
col = "blue", lwd =
2)
Data Exploration &
Plottings

CORRPLOT

library(corrplot)
corrplot(cor(
mtcars[,1:5]),method
="circle")

corrplot(cor(
mtcars[,1:5]),method
="numbers") #errors
make you learn

corrplot(cor(
mtcars[,1:5]),method
="square")
Save figure as Pdf

pdf(file = "My Plot.pdf",

width = 4, height = 4)
boxplot(mtcars[,1],
main="Miles/US gallon",
horizontal=T)

boxplot(mtcars[,3],
main="Displacement
(cu.in.)")
dev.off()

#or using loop

pdf(file = "My Plot

loop.pdf", width = 4,
height = 4)
for(i in 1:5){
boxplot(mtcars[,i],
main=paste("Column",i),
horizontal=T)
}
dev.off()
Exporting the dataset
Dataset write.table(mydata,
in "mydata.txt", sep=";")
text
file

Dataset
in
write.csv(mydata,
"mydata.csv") csv
file

Dataset
library(foreign)
in
write.dta(mydata,
Stata "mydata.dta")
file
 EXERCISE
1. Please import the data named “yourdata” using read.csv
command.

2. Give the mean of the hot variable in the yourdata.

3. What is the dimension of the yourdata dataset?

4. Which variables in the class of integer?

5. For the whole dataset, is there any observations that

uplicated?

6. Please create a command for variable wind that will give

sentence of “wind has duplicated values for more than half
of the dataset” if the variable contains more than 50
duplicated values. Otherwise, give “wind is okay”
EXERCISE
7. Extract the
output as the left
figure in your R
software.
Thank You!