data_exploration.r

#load data
data(iris)
#struktur dataset
str(iris)
#menampilkan beberapa data teratas
head(iris)
#membuat variabel iris.num untuk data numerik
iris.num <- iris[,1:4]

#statistika deskriptif
summary(iris)
#variansi atribut numerik
var(iris.num)

#membuat pie chart spesies
pie(table(iris$Species))

#mengatur layout 2x2, dengan outer margin 1
par(mfrow=c(2,2),oma=c(1,1,1,1))
#membuat histogram tiap atribut
hist(iris$Sepal.Length,col = "grey")
hist(iris$Sepal.Width,col = "grey")
hist(iris$Petal.Length,col = "grey")
hist(iris$Petal.Width,col = "grey")

#mengatur layout 2x2, dengan margin 2
par(mfrow=c(2,2),mar=c(2,2,2,2))
#membuat boxplot atribut terhadap spesies
boxplot(Sepal.Length ~ Species, main = "Box Plot Sepal Length - Species", data = iris, xlab = "Species", ylab = "Sepal.Length")
boxplot(Sepal.Width ~ Species, main = "Box Plot Sepal Width - Species", data = iris, xlab = "Species", ylab = "Sepal.Width")
boxplot(Petal.Length ~ Species, main = "Box Plot Petal Length - Species", data = iris, xlab = "Species", ylab = "Petal.Length")
boxplot(Petal.Length ~ Species, main = "Box Plot Petal Width - Species", data = iris, xlab = "Species", ylab = "Petal.Width")

#membuat pairs plot
iris.jitter <- apply(iris.num,2,function(o){jitter(o)})
pairs(iris.jitter,main="Pairs Plot of Iris Data",pch=21,bg = c("red", "green", "blue")[unclass(iris$Species)])