We can use the plot() function in the base plot system to create a scatter plot:
# Simply specify the x and y variables.
plot(pressure$temperature,pressure$pressure) 
Since there are only two variables, we can simply run:
plot(pressure)
pointslinestexts# Create the plot with title and axis labels.
plot(pressure,type="l",
main="Vapor Pressure of Mercury",
xlab="Temperature",
ylab="Vapor Pressure")
# Add points
points(pressure,col='red')
# Add annotation
text(150,700,"Source: Weast, R. C., ed. (1973) Handbook \n
of Chemistry and Physics. CRC Press.")
boxplot(hwy ~ cyl, data=mpg)
# Use the title function to add title and labels.
title("Highway Mileage per Gallon",
xlab = "Number of cylinders",
ylab = "Mileage (per gallon)")
hist() can used to create histogramshist(mpg$hwy)
hist(mpg$hwy, breaks=c(5,15,25,30,50))
The curve() function draws a function over a specified range.
curve(cos,-3*pi, 3*pi)
title("Cosine Function")
# The abline() function adds one or more straight lines to the current plot
abline(h=c(-1,0,1),
col = 2, lty = 2, lwd = 1.5) 
plot() without X and Y being specified, it will generate a panel grid of plots.str(airquality)## 'data.frame': 153 obs. of 6 variables:
## $ Ozone : int 41 36 12 18 NA 28 23 19 8 NA ...
## $ Solar.R: int 190 118 149 313 NA NA 299 99 19 194 ...
## $ Wind : num 7.4 8 12.6 11.5 14.3 14.9 8.6 13.8 20.1 8.6 ...
## $ Temp : int 67 72 74 62 56 66 65 59 61 69 ...
## $ Month : int 5 5 5 5 5 5 5 5 5 5 ...
## $ Day : int 1 2 3 4 5 6 7 8 9 10 ...plot(airquality)
Example: saving to a PNG file
png("test.png",width=5*240,height=3*240)
plot(pressure, type="l")
points(pressure,col="red")
dev.off()
qplot Functionqplot(weightLb, heightIn, data=heightweight, geom="point")
qplot(weightLb, heightIn, data=heightweight, geom ="text", label=ageYear)
Grammar of Graphics components:
ggplot function to indicate what data to usegeom_xxx functions to indicate what types of visual marks to use
aes() function) to map variables to visual marks
ggplot(heightweight, # What data to use
aes(x=weightLb,y=heightIn)) + # Aesthetic specifies variables
geom_point() # Geom specifies visual marks
This is equivalent to:
qplot(weightLb, heightIn, data=heightweight, geom="point")ggplot(mpg,aes(x=hwy)) + geom_bar(binwidth=5, fill="white", color="black")
ggplot(faithfuld, aes(waiting, eruptions, z = density))+
geom_raster(aes(fill = density)) +
geom_contour(colour = "white")
maps package, one can create geographical graphseast_asia <- map_data("world",
region=c("Japan","China","North Korea","South Korea"))
ggplot(east_asia, aes(x=long,y=lat,group=group, fill=region)) +
geom_polygon(color="black") +
scale_fill_brewer(palette="Set2")
geom_xxx functionsggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point() +
geom_quantile(quantiles = c(0.25,0.5,0.75)) +
geom_text(label=rownames(heightweight), vjust=-0.5)
Example: use the “sex” variable to group data points by shape and color:
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex))
Specify the shapes and colors manually:
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex),size=4) +
scale_shape_manual(values=c(1,4)) +
scale_color_manual(values=c("blue","green"))
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex,size=ageYear))
Use stat_smooth function to add a fitted model to the plot:
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex),size=4) +
scale_shape_manual(values=c(1,4)) +
scale_color_manual(values=c("blue","green")) +
stat_smooth(method = lm, level=0.95)
Moving the color=sex statement to the ggplot function produces two lines:
ggplot(heightweight, aes(x=weightLb,y=heightIn, color=sex)) +
geom_point(aes(shape=sex),size=4) +
scale_shape_manual(values=c(1,4)) +
scale_color_manual(values=c("blue","green")) +
stat_smooth(method = lm, level=0.95)
To label data points, use either annotate or geom_text
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex,size=ageYear)) +
annotate("text",x=150,y=68,label="Some label",color="darkgreen",size=12)
ggplot(heightweight, aes(x=weightLb,y=heightIn)) +
geom_point(aes(shape=sex,color=sex,size=ageYear)) +
geom_text(aes(label=ageYear),vjust=0.5)
stat_xxx function to choose a common transformation to visualize.We have seen the stat_smooth() function:
ggplot(heightweight, aes(x=weightLb,y=heightIn, color=sex)) +
geom_point(aes(shape=sex),size=4) +
scale_shape_manual(values=c(1,4)) +
scale_color_manual(values=c("blue","green")) +
stat_smooth(method = lm, level=0.95)
Another example: stat_bin() function creates a frequency count:
ggplot(mpg,aes(x=hwy)) + stat_bin(binwidth = 5)
This is equivalent to:
ggplot(mpg,aes(x=hwy)) + geom_bar(binwidth = 5)Or:
ggplot(mpg,aes(x=hwy)) +
geom_bar(stat="bin", binwidth = 5)
# The "bin" stat is the implied default for histogramDensity plot with stat_density:
ggplot(mpg,aes(x=hwy)) + stat_density()
Or the same plot with geom_bar:
ggplot(mpg,aes(x=hwy)) + geom_bar(stat="density")
ggplot2 plot can be saved in an objectp <- ggplot(heightweight, aes(x=weightLb,y=heightIn))
p + geom_point(aes(shape=sex,color=sex,size=ageYear))
# Here we use the saved plot object "p", but a different geom
p + geom_smooth(method=lm)
To add a title, use either ggtitle or labs(title=)
Note the title is left-aligned by default.
p <- ggplot(heightweight, aes(x=weightLb,y=heightIn, color=sex)) +
geom_point(aes(shape=sex),size=4) +
scale_shape_manual(values=c(1,4)) +
scale_color_manual(values=c("blue","green"))
p + ggtitle("Height ~ weight of school children")
xlab and ylab, orlabs(x=,y=)p + ggtitle("Height ~ weight of school children") +
xlab("Weight (lbs)") + ylab("Height (inch)")
labs(<aes>=) to specify legend titlesp + ggtitle("Height ~ weight of school children") +
xlab("Weight (lbs)") + ylab("Height (inch)") +
labs(color='Gender', shape='Gender')
guides function to set legend type for each aesthetic properties.Before:
p <- ggplot(heightweight, aes(x=weightLb,y=heightIn,color=ageYear)) +
geom_point(aes(shape=sex))
p
After:
p + guides(shape='none',color='legend')
ggthemes packageExample: Excel theme
p + theme_excel()
theme() function.
Example: removing the grid lines
p + theme_bw() +
theme(panel.grid = element_blank())
theme() function.
Or just removing the vertical ones:
p + theme_bw() +
theme(panel.grid.major.x = element_blank(),
panel.grid.minor.x = element_blank())
Change the base size and font family:
p + theme_bw(base_size = 24, base_family = "Times")
Or fine tune each element:
p + theme_bw(base_size = 24, base_family = "Times") +
theme(legend.title = element_text(size=20,color="blue"),# Legend title
legend.text = element_text(size=18,color="red"), # Legend text
axis.title.x = element_text(size=18,color="red"), # X axis label
axis.title.y = element_blank(), # Remove Y axis label
)
The element_blank() function can be used to remove undesired elements.
p + theme_bw(base_size = 24, base_family = "Times") +
theme(legend.position = "bottom")
p + theme_bw(base_size = 24, base_family = "Times") +
theme(legend.position = c(0.9,0.1))
theme_grey()theme_set() to change the defaultWith old default:
p
With new default:
theme_set(theme_light())
p
mytheme <- theme_bw(base_size = 24, base_family = "Times") +
theme(legend.title = element_text(size=20,color="blue"),# Legend title
legend.text = element_text(size=18,color="red"), # Legend text
axis.title.x = element_text(size=18,color="red"), # X axis label
axis.title.y = element_blank(), # Remove Y axis label
)
p + mytheme
g <- ggplot(mpg,aes(x=hwy)) +
geom_bar(binwidth=5, fill="white", color="black")Original:
g
With flipped coorinates:
g + coord_flip()
Original:
g
With transformed Y coordinate:
g + coord_trans(y="sqrt")
scale_x_continuous: scale for X, which is a continuous variable
p + theme_bw() +
ylim(50,100) +
scale_x_continuous(limits=c(0,200),
breaks=c(50,110,170),
labels=c("Thin","Medium\nSize","Chubby"))
scale_x_log10: base-10 logarithmic scale for X, which is a continuous variable
p + theme_economist_white() +
scale_x_log10(breaks=c(10,20,50,100,200),
limits=c(5,500)) + # Plot X on a log10 scale
scale_y_reverse() # Reverse the Y scale
Original Boxplot:
ggplot(mpg,aes(x=drv,y=cty,fill=drv)) +
geom_boxplot()
Use scale_fill_discrete function to modify legend labels:
By default:
ggplot(mpg,aes(x=displ,y=hwy,size=cyl,color=drv, alpha=cty)) +
geom_point()
Re-scaled:
ggplot(mpg,aes(x=displ,y=hwy,size=cyl,color=drv, alpha=cty)) +
geom_point() +
scale_size_identity() + # Use the values of "cyl" variable for size
scale_color_manual(values=c("darkblue","rosybrown2","#24FA22")) +
scale_alpha_continuous(range=c(0.1,1))
ggplot2 is managed by the functions facet_grid and facet_wrap.facet_grid: create a row of panels defined by the variable “drv”:
ggplot(mpg, aes(displ, hwy)) +
geom_point() +
facet_grid(. ~ drv)
facet_grid: creates a column of panels defined by the variable “fl”:
ggplot(mpg, aes(displ, hwy)) +
geom_point() +
facet_grid(fl ~ .)
facet_grid: creates a matrix of panels defined by the variables “fl” and “drv”:
ggplot(mpg, aes(displ, hwy)) +
geom_point() +
facet_grid(fl ~ drv)
facet_wrap: wraps 1d sequence of panels into 2d:
ggplot(mpg, aes(displ, hwy)) +
geom_point() +
facet_wrap(~class, nrow=3)