Visualisation

Packages often used are MASS and ggplot2, and in 3142 we also use ISLR

library('ggplot2')
library('MASS')
library('ISLR')

## Warning: package 'ISLR' was built under R version 3.6.3

Base R plot examples

Bar plot Example:

obs_bar <- rnorm(100,3,2)
hist(obs_bar, xlab = "Observations")

Line Graph Example:

obs_line <- c(rnorm(500, 3, 1), rnorm(500,3,3))+(1:1000)/100
plot(1:length(obs_line),obs_line,"l", xlab = "Time", ylab = "Value")

Scatter plot Example:

plot(mammals$body, mammals$brain)

Boxplot Example:

boxplot(decrease ~ treatment, data = OrchardSprays,log = "y", col = "bisque")

Further Line Graph Example:

plot(longley$Year, longley$Armed.Forces, type="l", lwd = 2, main="Armed Forces and Unemployment (1947-1962)", sub="Exploratory analysis",
     xlab="Year", ylab="Number", ylim=c(0,600))
points(longley$Year, longley$Unemployed, type="l", lwd = 2, col="red") #Same as plot but does not create an entirely new plot but rather adds onto the original plot
legend('bottomright', legend=c("Armed Forces","Unemployed"), col=c("black","red"), lty=c(1,1))

Curve Function Example:

curve(dnorm(x, 0, 1/3), from=-2, to=2, lwd = 2)

Advanced curve example:

set.seed(1)
hist(rnorm(10000, -3, 1), breaks = 50, prob = T, col = "lightblue", xlab = "", xlim = c(-8, 8), main = "Two empirical distributions")
hist(rnorm(10000, 3, 1), breaks = 50, prob = T, col = "blue", add =T) #add=T makes it add ontop of the created plot
curve(dnorm(x, -3, 1), lwd = 3, add = T, col = "red")
curve(dnorm(x, 3, 1), lwd = 3, add = T, col = "red")
legend("topright", ncol = 1, cex = 1.5, legend=c("Sample 1", "Sample 2"), fill=c("lightblue", "blue"))

Graphical windows example

# Split the window to have 3 lines and 2 columns of plots (to be filled by column)
par(mfcol=c(3,2))
# First plot
plot(1:3,1:3,col="red",type="l", xlab= "", ylab= "")    
plot(1:5,1:5,col="orange",type="p", xlab= "", ylab= "") 
plot(1:7,1:7,col="blue",type="b", xlab= "", ylab= "")   
plot(ts(rnorm(100)),col="purple", xlab= "", ylab= "")
curve(sqrt(x^2 - x^4), from = -1, to = 1, lwd = 2, xlab= "", ylab= "")
hist(rexp(100,1/2), main = "", xlab= "", ylab= "")

You can add lines to a plot manually with the abline function. It can be in the form \(y=a+bx\) or in terms of horizontal and vertical location:

plot(0,0,"n", xlab="", ylab="") 
abline(h=0.5, v=0, col = "purple", lwd = 4)
abline(a=1, b=1, col = "orange", lwd = 4)

Can add lines through the use of segments which join set coordinates together:

# create an empty plot
plot(0,0,"n", xlab="", ylab="") 
# add a segment joining two points
segments(x0=-0.5, y0=0, x1=1, y1=-0.5, col = "red", lwd = 4)
# add two segments, joining three points: first argument is the 'x' values, second is the 'y' values
lines(c(-1,0,1), c(-1,0.2,1), col = "blue", lwd = 4) 

Contour Plots can be used to create more sophisticated plots. They can be used to represent three dimensional data, like a topographical map. The contour() function takes a x vector, a y vector and a matrix whose elements respond to the z value for each pair of x,y coordinations

x = seq(-pi, pi, length=50)
y = x
f = outer(x,y, function(x,y) cos(y)/(1+x^2))
contour(x,y,f)
contour(x,y,f,nlevels=45,add = T)

The pairs() function can be used to create a scatterplot matrix. This can also be done for just a subset of variables

pairs(Auto) #For all variables

pairs(~ mpg + displacement + weight + acceleration, Auto)

ggplot2 Graphs

ggplot is a layering system, where you add layers of visulation ontop of the prior * Needs data to be in a data frame * Use aes() in the first ggplot statement if is the same for all plots * aes stands for aesthetic and is telling R what we want to see on the graph Examples:

data <- as.data.frame(Seatbelts)
ggplot(data) +
  geom_histogram(aes(drivers), bins = 10) +
  labs(x = "monthly car drivers killed or seriously injured", 
       title = "Histogram of  UK - Jan 1969 to Dec 1984")

data$law <- factor(data$law,  labels = c("0", "1"))
ggplot(data) +
  geom_boxplot(aes(x = law, y = drivers)) +
  labs(x = "Mandatory Seatbelts: NO(0) or YES(1)", 
       y = "Deaths & Serious Injuries")

ggplot(mpg, aes(cty, hwy)) + 
  geom_point()+ 
  geom_smooth(method="lm", se=T) + 
  theme_bw()

Can use stat functions with ggplot aswell, which allows the plotting of a function:

ggplot(data = data.frame(x = c(-3, 3)), aes(x)) +
  stat_function(fun = dnorm, n = 100, args = list(mean = 0, sd = 1))

ggplot(midwest, aes(x=area, y=poptotal)) + 
  geom_point(aes(col=state, size=popdensity)) + 
  xlim(c(0, 0.1)) + 
  ylim(c(0, 500000)) + 
    theme(text = element_text(size=20)) #make axis bigger

## Warning: Removed 15 rows containing missing values (geom_point).