Skip to main content

Community Contributions STAT GR 5293

25 Drawing Five Common Plots by ggplot2

Xirui Guo

25.1 Motivation

The ggplot2 is a powerful and effective package for drawing plots. It has a lot of syntaxes to support people get the graphs they want. A good plot needs to clearly represent the information. Scatter plots, line plots, histograms, boxplots, and heatmaps are frequently used in daily life. This Markdown hopes to be a guide about how to quickly and suitably draw the above five plots by ggplot2.

25.2 Graph1 scatter plot: geom_point

Using data mtcars
The main code for drawing scatter plot is geom_point; however, we usually don’t use it single because we need the plot shows more information and here are some syntaxes people use together with geom_point:
aes(size , color): sometimes it is effective, but people need to avoid showing too much information and making the graph hard to read..
alpha: change the transparency of point.
geom_text: add labels of each point.
geom_smooth: show the regression line and correspond standard error boundary.

mtcars$vs <- as.factor(mtcars$vs)

ggplot(mtcars, aes(x=wt, y=mpg, size=cyl, color=vs)) +
  geom_point(alpha = .4) + 
  geom_text(label=rownames(mtcars), 
            nudge_x = 0.25, nudge_y = 0.25, 
            check_overlap = T)+
  geom_smooth(method=lm , color="red", fill="lightblue", se=TRUE)

25.3 Graph2 line chart: geom_line

People need to clarify the x-axis and y-axis and can use group to draw the line with different colors.

# library(reshape2)
x <- -10:10
var1 <- dnorm(x,-2,1)
var2 <- dnorm(x,2,1)
var3 <- dt(x,2,2)
data <- data.frame(x,var1,var2,var3)
data_n <- melt(data, id="x")

ggplot(data_n,aes(x=x, y=value, group=variable, color=variable))+
  geom_line()

Using geom_line twice: a way for representing line from different groups without legend.

data_n <- data_n %>%
  mutate(variable2=variable) 

ggplot(data_n, aes(x=x, y=value))+
  geom_line(data=data_n %>% select(-variable), aes(group=variable2), color="grey", size=0.5, alpha=0.8) +
  geom_line(aes(color=variable), color="blue", size=1)+
  facet_wrap(~variable)

25.4 Graph3 box plot: geom_boxplot

Using the Data: StudentSurvey in Lock5withR package(PSet1). Drawing a “simple” box-plot first:
labs(): simultaneously add the main title and axis labels

# library(Lock5withR)
p3<-ggplot(StudentSurvey, aes(x = SAT, y = Year)) + 
  geom_boxplot()+
  labs(title="Plot of SAT score and Year", x="SAT score", y="Year")
p3

How to make boxplot shows more information:
1. Using theme() + element_text() to modify the color, font, size of text.

p3+theme(
  plot.title = element_text(color = "red", size = 14, face = "bold.italic"),
  axis.title.x = element_text(color="blue", size = 14, face = "bold"),
  axis.title.y = element_text(color="green", size = 14, face = "bold")
)
  1. Adding legend and modifying the legend tittle
p3<-ggplot(StudentSurvey, aes(x = SAT, y = Year, fill=Year)) + 
  geom_boxplot()+
  labs(fill="School Year")
p3

3.Changing the position of legend: five choices “left”,“top”, “right”, “bottom”, “none”

p3<- p3+
  theme(legend.position = "top")
p3
  1. Modifying the title, label and background of legend
p3+
  theme(
    legend.title = element_text(color="blue"),
    legend.text = element_text(color="red"),
    legend.background = element_rect(fill="lightblue")
    )
  1. Changing the name and order of factor’s level
Y <- fct_recode(StudentSurvey$Year, "NA" = "", "1st" = "FirstYear", 
                "2nd" = "Sophomore", "3rd" = "Junior", "4th" = "Senior")
Y <- fct_relevel(Y,"1st","2nd","3rd","4th")

p3<-ggplot(StudentSurvey, aes(x = SAT, y = Y, fill=Y)) + 
  geom_boxplot()+
  labs(fill="School Year")
p3
  1. 6 ways to change the background of whole plots
p3_1<-p3+theme_gray()
p3_2<-p3+theme_bw()
p3_3<-p3+theme_linedraw()
p3_4<-p3+theme_light()
p3_5<-p3+theme_minimal()
p3_6<-p3+theme_classic()

ggarrange(p3_1,p3_2,p3_3,p3_4,p3_5,p3_6,nrow=3,ncol=2)

25.5 Graph4 histogram: geom_histogram

Using Data: bad_drivers in fivethirtyeight package. Common syntaxes with geom_histogram:
binwidth: control the width of each bin.
fill: the color of bin.
color: the frame color of bin.
alpha: control the transparency of the bin color.

# library(fivethirtyeight)
p4 <- ggplot(bad_drivers,aes(x=perc_alcohol,y=..count..)) +
  geom_histogram(binwidth=3, fill="lightblue", color="black", alpha=0.7)
p4

coord_flip()：make the histogram bar become horizontal

p4<- p4 +
  coord_flip()
p4

scale_x_reverse() and scale_y_reverse()：reverse the x-axis and y-axis

25.6 Graph5 heatmap: geom_tile

The data is from the class notes

grade <- rep(c("first", "second", "third", "fourth"), 3)
subject <- rep(c("math", "reading", "gym"), each = 4)
statescore <- sample(50, 12) + 50
df <- data.frame(grade, subject, statescore)

p5 <- ggplot(df, aes(grade, subject, fill = statescore)) + 
  geom_tile() 
p5
  1. Changing the color scale_fill_gradient(), scale_fill_distiller() and scale_fill_viridis()
    scale_fill_gradient() can customize the color
    scale_fill_distiller() usually need palette = RColorBrewer
    scale_fill_viridis() need to let discrete=False when the variable is continuous
# library(RColorBrewer)
display.brewer.all()
p5_1<-p5 + scale_fill_gradient(low="white", high="purple") + theme(legend.position="top")
p5_2<-p5 + scale_fill_distiller(palette = "RdBu")+ theme(legend.position="top")
p5_3<-p5 + scale_fill_viridis_c()+ theme(legend.position="top")
ggarrange(p5_1,p5_2,p5_3,nrow=1,ncol=3)
  1. adding text in each square and interact
# library(plotly)

df <- df %>%
  mutate(text = paste("grade ", grade , "\n", "subject: ", subject, "\n", "statescore: ",statescore))
pp5<-ggplot(df, aes(grade, subject, fill = statescore, text = text)) + 
  geom_tile() 
ggplotly(pp5, tooltip="text")

25.7 Effective ways for showing more information

25.7.1 Way1: Faceting

  1. faceting by single discrete variable：
# vertical faceting
p3 + 
  facet_grid(Award~.)
# horizontal faceting
p3 + 
  facet_grid(.~Award)
  1. faceting by two discrete variables：
# column facet by Year and row facet by Award
p3+
  facet_grid(Year~Award, scales='free')

25.7.2 Way2: Representing multiple charts on a single page

Using Data: bad_drivers with syntaxes grid.arrange()

# Make 3 simple graphics:
g6_1 <- ggplot(mtcars, aes(x=qsec)) + 
  geom_density(fill="slateblue")

g6_2 <- ggplot(mtcars, aes(x=drat, y=qsec, color=cyl)) + 
  geom_point(size=5) + theme(legend.position="none")

g6_3 <- ggplot(mtcars, aes(x=factor(cyl), y=qsec, fill=cyl)) + 
  geom_boxplot() +
  theme(legend.position="none")

# Plots
grid.arrange(g6_1, arrangeGrob(g6_2, g6_3, ncol=2), nrow = 2)
grid.arrange(g6_1, g6_2, g6_3, nrow = 3)