ggplot2 in Action - Part 2 (Update soon)
Written on November 3, 2016 by hoanguc3m
I really like the default R plot system. It always brings elegant minimal clean plots which suits with most of my purpose. However, there would be some circumstance that the default plot is way too simple and could not help me explain the data concretely. This website here is an awesome source of reference which convinces that you could do everything with R default plot (and even more).
How about elegant figures with elegant commands. Cool eh,…
These plots are rendered using ggplot2 - “the Grammar of Graphics”. I borrow the section names in shinyapps.org
Correlations
ggplot(cor_mat, aes(x=height.ratio,y = pop.vote)) +
geom_point(size = 5,shape = 21, colour = "black", fill = "grey") +
geom_smooth(method = "lm", se = F) +
xlab("Presidential Height Ratio") +
ylab("Relative Support for President") +
annotate("text", x = 1.15, y = .61, label = "r = .39") +
scale_y_continuous(breaks=seq(0.40, 0.7, 0.05)) +
scale_x_continuous(breaks=seq(0.9, 1.2, 0.05)) +
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5))
Histograms
Including “rug” Tick Marks
ggplot(good.choices.df, aes(x=good.choices)) +
geom_histogram(breaks=seq(0.3, 0.8, 0.05), colour = "black", alpha = 0.3) +
geom_rug(sides = "b", aes(y = 0), position = "jitter", colour = 2) +
xlab("Prop. Choices from Good Decks") +
ylab("Number of Studies") +
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5))
Including a Density Estimator
ggplot(good.choices.df, aes(x=good.choices,..density..)) +
geom_histogram(breaks=seq(0.3, 0.8, 0.05), colour = "black", alpha = 0.3) +
geom_rug(sides = "b", aes(y = 0), position = "jitter", colour = 2) +
geom_density(colour = "black") +
xlab("Prop. Choices from Good Decks") +
ylab("Density of Studies") +
xlim(0.3, 0.8) +
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5))
Including Numbers on Top
ggplot(good.choices.df, aes(x=good.choices)) +
geom_histogram(breaks=seq(0.3, 0.8, 0.05), colour = "black", alpha = 0.3) +
stat_bin(breaks=seq(0.4, 0.7, 0.05), geom="text", aes(label=..count..), vjust=-0.5) +
geom_rug(sides = "b", aes(y = 0), position = "jitter", colour = 2) +
xlab("Prop. Choices from Good Decks") +
ylab("Number of Studies") +
xlim(0.3, 0.8) +
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5))
Line Plots
Regular Line Plot
ggplot(RT, aes(x=word_frequency, class = AS)) +
geom_point(aes(y = mean, shape = AS), size = 4) +
geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.1) +
geom_line(aes(y = mean, group = AS),linemitre = 100, linejoin = "bevel") +
xlab("Word Frequency") +
ylab(expression(paste("Mean ", mu))) +
scale_y_continuous(breaks=seq(0.30, 0.6, 0.05),limits =c(0.3, 0.6) ) +
scale_shape_manual(values=c(1, 19))+
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position = c(0,1), legend.justification = c(-0.5, 1),
legend.title=element_blank()) 
Box Plot
ggplot(data = RT, aes(x=word_frequency,y = mean, class = AS)) +
geom_boxplot(aes(colour = AS), outlier.size = 0,width = 0.3, position = "identity", fill = "white") +
stat_summary(geom = "point",aes(colour = AS), fun.y = "mean", size = 3) +
stat_boxplot(aes(colour = AS), geom ='errorbar', width = 0.3, position = "identity") +
xlab("Word Frequency") +
ylab("Group Mean M") +
scale_y_continuous(breaks=seq(0.30, 0.6, 0.05),limits =c(0.3, 0.6) ) +
scale_colour_manual(values=c("gray","black"))+
scale_alpha_manual(values=c(0.5,1))+
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position = "none") +
annotate("text", x = 2, y = 0.56, label = "Accuracy", color="gray", alpha = 1) +
annotate("text", x = 2, y = 0.35, label = "Speed", color="black", alpha = 1)
Violin Plot
ggplot(data = RT, aes(x=word_frequency,y = mean, class = AS)) +
geom_violin(aes(colour = AS, fill = AS), width = 0.3, position = "identity") +
stat_summary(geom = "point", fun.y = "mean", size = 3, shape = 13) +
stat_boxplot(aes(colour = AS), geom ='errorbar', width = 0.3, position = "identity") +
xlab("Word Frequency") +
ylab("Group Mean M") +
scale_y_continuous(breaks=seq(0.30, 0.6, 0.05),limits =c(0.3, 0.6) ) +
scale_colour_manual(values=c("gray","black"))+
scale_fill_manual(values=c("gray","white"))+
scale_alpha_manual(values=c(0.5,1))+
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position = "none") +
annotate("text", x = 2, y = 0.56, label = "Accuracy", color="gray", alpha = 1) +
annotate("text", x = 2, y = 0.35, label = "Speed", color="black", alpha = 1)
Combined Line and Bar Plot
Well, I had some problems with this graph. The object is to plot 2 different scales on the left and the right hand sides. This type of graph usually makes people confused at first and requires more effort.
Well, my first solution here is to rescale data to fit with one range. The other solution is dual axis (see rpubs) where we plot them separately and combine into a final plot.
In comparison to the original R plot, ggplot2 makes a great improvement and easy maintenance. Es fantastico!
ggplot(data = RTER.MRT, aes(x=cue,y = mrt)) +
geom_point(size = 4) +
geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.1) +
geom_line(aes(y = mrt, group = 1)) +
geom_text(aes(x=cue, y = mrt+50 ,label = mrt) ) +
geom_errorbar(data = RTER.Er, aes(ymin = lower*500 + 150, ymax = upper*500+150), width = 0.2) +
geom_bar(data = RTER.Er.Mod,aes(y = mrt), stat = "identity", colour ="black", fill = "white", width = 0.5) +
scale_x_discrete(limits=c("Speed","Neutral","Accuracy")) +
xlab("Cue") +
ylab("Mean Response Time (ms.)") +
ggtitle("Behaviour Data") +
geom_text(data = RTER.Er.Mod, aes(x=cue, y = mrt - 50,label = RTER.Er$mrt) ) +
scale_y_continuous(breaks=seq(300, 700, 100),limits =c(000, 700) ) +
theme_classic()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position = "none") 
p1 <- ggplot(data = RTER.MRT, aes(x=cue,y = mrt)) +
geom_point(size = 4) +
geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.1) +
geom_line(aes(y = mrt, group = 1)) +
xlab("Cue") +
ylab("Mean Response Time (ms.)") +
ggtitle("Behaviour Data") +
geom_text(aes(x=cue, y = mrt+50 ,label = mrt) ) +
scale_x_discrete(limits=c("Speed","Neutral","Accuracy")) +
scale_y_continuous(breaks=seq(300, 700, 100),limits =c(200, 700) ) +
theme_bw()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
legend.position = "none",
plot.margin = unit(c(1,1.8,1,1), "cm") )
p2 <- ggplot(data = RTER.Er, aes(x=cue,y = mrt)) +
geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.2) +
geom_bar(stat = "identity",width = 0.5, colour ="black", fill = "white") +
xlab("Cue") +
ylab("Mean Proportion of Errors") +
geom_text(data = RTER.Er, aes(x=cue, y = mrt-0.05 ,label = mrt) ) +
scale_x_discrete(limits=c("Speed","Neutral","Accuracy")) +
scale_y_continuous(breaks=seq(0, 0.4, 0.1),limits =c(0,1.2) ) +
theme_bw()+
theme(axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
legend.position = "none",
plot.margin = unit(c(1,2,1,1), "cm"))
p <- ggplot_dual_axis(lhs = p1, rhs = p2)
grid.draw(p)