Understanding GGPLOT function

Simple plot showing city Mileage Vs Displacement by Model

p <- ggplot(data=mpg, aes(x=displ, y=cty, color=model))
p + geom_point(size=4)

Simple plot showing city Mileage Vs Displacement by Fuel type

p1 <- ggplot(data=mpg, aes(x=displ, y=cty, color=fl))
p1 + geom_point(size=4)

Simple plot showing city Mileage Vs Displacement by class of vehicle

p2 <- ggplot(data=mpg, aes(x=displ, y=cty, color=class))
p2 + geom_point(size=4)

Adding a geom Layer

p <- ggplot(data=mpg, aes(x=displ, y=cty, color=class))
p+geom_blank()

p+geom_point()

p+geom_jitter()

p+geom_count()

Connecting points in the plot
```
p+geom_line()
```
```
p+geom_step()
```
```
p+geom_path()
```

Displaying data and statistical summary

p <- ggplot(data=mpg, aes(x=displ, y=cty))
p+geom_point() + geom_smooth() # Smooth fit line

## `geom_smooth()` using method = 'loess'

p+geom_point() + geom_smooth(method = "lm") # Regression line

Displaying Bar graphs

p<- ggplot(data=mpg, aes(x=class))
p+geom_bar()

p<-ggplot(data=mpg,aes(x=model))
p+geom_bar(color="blue", fill="light blue")

Constructing box plots

p <- ggplot(data=mpg, aes(x=class, y=cty))
p+geom_boxplot()

p+geom_violin()

Density plot using ggplot

p <- ggplot(data=mpg, aes(x=cty))
p+geom_density(fill="light blue")

Histogram plot using ggplot

p <- ggplot(data=mpg, aes(x=cty))
p+geom_histogram(binwidth = 1.8, color="blue", fill="light blue")

Add or remove aesthetic mapping

Add aesthetic mapping

p<-ggplot(data=mpg,aes(x=displ, y=cty, color=fl))
p + geom_point(aes(shape=fl)) + geom_smooth(method = "lm", se=FALSE)

Remove aesthetic mapping

p<-ggplot(data=mpg,aes(x=displ, y=cty, color=fl))
p + geom_point(aes(color=NULL)) + geom_smooth(method = "lm", se=FALSE)

parameter setting Vs Aesthetic mapping

p<-ggplot(data=mpg,aes(x=displ, y=cty))
p + geom_point(color="red") # color Parameter set to red

p + geom_point(aes(color=fl)) # Aesthetic mapping

Adjusting the positions of Bars in the bar plot

p <- ggplot(data=mpg, aes(x=cty, fill=class))
p + geom_bar()

p + geom_bar(position="stack")

p + geom_bar(position="dodge")

p + geom_bar(position="fill")

Adjusting the width of the Bars in the bar plot

p <- ggplot(data=mpg, aes(x=cty))
p + geom_bar()

p + geom_bar(width = 0.9) # default

p + geom_bar(width = 0.95)

p + geom_bar(width = 0.5)

Adjusting the position of points

p <- ggplot(data=mpg, aes(x=displ, y=cty))
p + geom_point()

p + geom_jitter()

Adjusting symbol size shape jitter and transparency

p <- ggplot(data=mpg, aes(x=displ, y=cty))
p + geom_point(size = 3, alpha=1/2, color="red")

p + geom_point(size = 3, alpha=1/10, color="red")

p + geom_jitter(size = 4, alpha=1/8, color="red")

p + geom_point(shape=2, size = 3, alpha=1/2, color="red")

p + geom_point(shape=4, size = 2, alpha=1/5, color="red")

Changing the coordinate system

p <- ggplot(data=mpg, aes(x=factor(1), fill=class))
p + geom_bar()

# Polar coordinate system
p + geom_bar() + coord_polar(theta = "y")

# flipping coordinates
p + geom_bar() + coord_flip()

p + geom_bar() + coord_polar(theta = "y", direction = -1)

Each plot layer may contain data from different dataframes

mpg_midsize <- subset(mpg, class == "midsize")
mpg_pickup <- subset(mpg, class == "pickup")
mpg_suv <- subset(mpg, class == "suv")

# Combine plots of two data frames
p <- ggplot(data = mpg_midsize, aes(x=displ, y=cty))
p + geom_jitter(color="blue") + geom_jitter(data = mpg_suv, color="red") + geom_jitter(data = mpg_pickup, color="green")

mpg_midsize_suv <- rbind(mpg_midsize,mpg_suv)

# Plotting combined data 
p <- ggplot(data = mpg_midsize_suv, aes(x=displ, y=cty, color=class))
p + geom_jitter()

Adding Labels to your plot: Labeling points on scatter plot

p <- ggplot(data=mpg_midsize, aes(x=displ,y=cty))
p + geom_point(size=3) + geom_text(aes(label=model), nudge_y = 0.4, nudge_x =0.12 , size=3) + xlim(2,9) + ylim(6, 36)

## Warning: Removed 2 rows containing missing values (geom_point).

## Warning: Removed 2 rows containing missing values (geom_text).

p + geom_point(size=3) + geom_label(aes(label=model), nudge_x = 0.12,nudge_y = 0.3)

Adding title to the plot

 p + geom_point(size=3) + geom_label(aes(label=model), nudge_x = 0.12,nudge_y = 0.3) + ggtitle("City Mileage Vs Displacement")

# Non overlapping labels
p <- ggplot(data=mpg_midsize, aes(x=displ,y=cty))
p + geom_point(size=3) + geom_label(aes(label=model), label.size = 0.2) # need to check !

Using Scale functions

p <- ggplot(data=mpg, aes(x=cty, fill=class))
p + geom_bar(color="black") + scale_fill_discrete()

p + geom_bar(color="black") + scale_fill_brewer()

choosing a color palette

library(RColorBrewer)
display.brewer.all()

p <- ggplot(data=mpg, aes(x=cty, fill=class))
p + geom_bar(color="black") + scale_fill_brewer(palette = "Accent")

p + geom_bar(color="black") + scale_fill_brewer(palette = "Spectral")

Position Scales

p <- ggplot(data=mpg, aes(x=displ,y=cty))
p + geom_jitter()

# reversing the y scale
p + geom_jitter() + scale_y_reverse()

# changing the y-axis scale to log10
p + geom_jitter() + scale_y_log10()

# setting y axis labels
p + geom_jitter() + scale_y_log10(breaks=c(0.5, 8, 20), labels=c(0.5,8,20))

Theme: Titles, Tick Marks, and Tick Labels

p <- ggplot(data=mpg, aes(x=displ,y=cty))
q <- p + geom_jitter() + ggtitle("City Mileage Vs Displacement") + xlab("Engine Displacement") + ylab("City Mileage")+scale_x_continuous(breaks = seq(1.2,7, by=0.25)) 
q + theme(plot.title = element_text(colour = "blue", size = 20, hjust=0.5), axis.title = element_text(color = "blue" , size = 10))

Using Facets

p <- ggplot(data=mpg, aes(x=displ,y=cty))
p + geom_jitter() + facet_grid(.~class)

Understanding GGPLOT function

Milind DESAI

October 19, 2018