Graphics

Base Statistical Graphics

Univarite

========================================================

Histogram, Stem-and-Leaf, Dot, Q-Q, Density plots

#Histogram
x = DataX$Sepal.Length # a continuous variable
by (mfrow = c (2,2))
hist(x, main='Histogram (Default)')   
hist(x, breaks=20, main='More bins')  
hist(x, breaks=20, freq=F, main='Frequency vs. Density')  
hist(x, breaks=20, freq=F, col=5, main='Coloring') 

#Density Plot
hist(x, breaks=20, freq=F,  
     main='Histogram with Kernel Density Plot')   # using freq=FALSE
lines(density(x), col=2, lty=1, lwd=2)  #add the density curve

Boxplot, Box-and-whisker

#Boxplot Plot
by (mfrow = c (2,2))
boxplot(DataX$Sepal.Length, main='Boxplot of Sepal.Length') 
boxplot(DataX$Sepal.Width, main='Boxplot of Sepal.Width')  # Outliers
boxplot(DataX[,1:4], col=c(2,3,4,5), main='Side-by-side Boxplot')  
boxplot(Sepal.Width~Species, DataX, col=c(6,7,8), main="Boxplot with Grouping") 

Bar, Pie, Polar, Waterfall charts

#Pie and Bar Charts
DataX$Flag = DataX$Sepal.Length>5 # Create a binary flag
by (mfrow = c (1,3))
pie(table(DataX$Species[DataX$Flag]), col=c(2,3,4))
barplot(table(DataX$Species[DataX$Flag]), col=c(5,6,7))
barplot(table(DataX$Species, DataX$Flag), col=c(5,6,7), beside=T)

Relationship between variables

x = DataX$Petal.Length; y = DataX$Petal.Width
plot(x, y, xlab="Petal.Length", ylab="Petal.Width") 
abline(lm(y~x), col=1, lty=2)
title(main=paste("Correlation =", round(cor(x,y),4)))

Differentiating with colors

z = DataX$Species
plot(x, y, col=c(2,3,4)[z], pch=20, cex=2.0, xlab="Petal.Length", ylab="Petal.Width") 
abline(lm(y~x), col=1, lty=2) 
legend("topleft", levels(z), pch=20, col=c(2,3,4), cex=2.0)

Distribution

library(lattice)
library(gridExtra)
p1 = histogram(DataX$Sepal.Length)
p2 = bwplot(DataX$Sepal.Length) 
grid.arrange(p1, p2, ncol=2)
#Histogram with Conditioning
histogram(data=DataX, ~Sepal.Length|Species, breaks=12, layout = c(3, 1))
#Density plot with Grouping
densityplot(data=DataX, ~Sepal.Length, groups=Species,  
            plot.points=F, auto.key=list(space="top", columns=3))
#Boxplot with Grouping
bwplot(data=DataX, Sepal.Width~Species) 

Bivariate

XYplot, Line, Area, Scatter, Bubble charts

#Pairwise Scatter Plot
pairs(DataX[,1:4], main="Pairwise Scatter Plots (default)") 
#Pairwise Scatter Plot (more sophisticated)
pairs(DataX[,1:4], panel = panel.smooth, 
      pch=21, bg = c(5,6,7)[unclass(DataX$Species)], 
      main="Pairwise Scatter Plot (sophisticated)") 

Distribution

#Grouping
xyplot(data=DataX, Sepal.Length ~ Petal.Length, groups = Species,  
       type = c("p", "smooth",  "g"),  
       auto.key = list(space="top", columns=3)) # grouping
#Conditioning
xyplot(data=DataX, Sepal.Length ~ Petal.Length|Species, 
       type=c("p", "smooth",  "g"), layout=c(1,3)) # conditioning

Trivariate

3D Scatter, Contour, Level/Heatmap, Surface plots

#3D Plot
cloud(data=DataX, Sepal.Length ~ Sepal.Length * Petal.Width, groups = Species, 
      auto.key = list(space="top", columns=3), panel.aspect = 0.8)
#Heatmap 
dist = as.matrix(dist(DataX[,3:4]))
levelplot(dist, colorkey = T, col.regions = terrain.colors,
          scales = list(at=c(0,0),tck = c(0,0)), 
          xlab="",ylab="",main="Levelplot of Pairwise Distance Matrix")