1 Intro

This is my guide to ggplot2 and related packages. It contains:

Web sites, blogs and books for ggplot2
Some examples about the basics of ggplot2:

Basic plots
Basic function of ggplot2(legends, titles etc)

Some advanced examples of more complex visualizations
R version :

R.Version()$version.string

## [1] "R version 3.4.1 (2017-06-30)"

2 Sources, Tutorials, examples and books

This site has most of ggplot2 functions. It is the ggplot2 website(part of tidyverse) with many examples.
Package ggthemes has many themes for ggplot2. Here you can see most of ggthemes functions with examples. It may be introductory but it helps to pick a theme.
R Graphics Cookbook: A very useful book, not only to begin with but to have as a permanent guide. It contains multiplot() function(uses package grid, in base package in R 3.4.1).
RcolorBrewer contains beautifull palletes but most usefull is the addin for R-studio Colourpicker. You can pick your colour in a new window and it returns the related ggplot2 code in your script.
Top 50 ggplot2 Visualizations gives some examples of more complex and beautifull plots for many types of data.
ggtern is an extension to ggplot2 specifically for the plotting of ternary diagrams.
Othes packages to combine with: library(ggcorrplot); library(ggExtra); library(scales); library(ggalt); library(ggfortify); library(ggdendro); library(ggmap); library(gridExtra)' ';ggbiplot; require(GGally); require(ggvis)
Useful R studio Addins: Colourpicker, Marginal Plots, Plot colour helper

3 Datasets and libraries

Lets have a look to the datasets I use to demonstrate ggplot2 plots.
At this point, I want to point out the “df_print: true” option so the html output shows a data frame object as a table with multiple pages.

require(ggplot2) 
require(RColorBrewer)  # many nice palettes
require(dplyr)  # data wrangling
require(ggthemes)  # additional themes 
require(knitr)     
require(xtable)   #for tables
library(ggcorrplot)    # corrplots
library(ggExtra)
library(scales)
library(ggalt)
library(ggfortify)
library(gridExtra)   # Multiple plots in one page and more
library(ggdendro)    # dendrogram, plot clusters and tree based methods
library(ggmap)       # maps
require(GGally)
require(ggvis)
require(ggpubr)     # Add Statistical test results into the plot(compare means etc), arrange plots

options(xtable.comment=FALSE)# This removes the comment % above xtable
print(xtable(summary(iris),
  digits=2,
  align = "cccccc"),
  include.rownames=F)# include.rownames=F removes the first column which is just a count

kable(summary(iris)) # for html table output to demonstrate this way to tables too

Sepal.Length	Sepal.Width	Petal.Length	Petal.Width	Species
Min. :4.300	Min. :2.000	Min. :1.000	Min. :0.100	setosa :50
1st Qu.:5.100	1st Qu.:2.800	1st Qu.:1.600	1st Qu.:0.300	versicolor:50
Median :5.800	Median :3.000	Median :4.350	Median :1.300	virginica :50
Mean :5.843	Mean :3.057	Mean :3.758	Mean :1.199	NA
3rd Qu.:6.400	3rd Qu.:3.300	3rd Qu.:5.100	3rd Qu.:1.800	NA
Max. :7.900	Max. :4.400	Max. :6.900	Max. :2.500	NA

mtcars

diamonds

4 Multiple plots in one page: Multiplot Function, gridExtra, ggpubr

Fit multiple ggplot2 plots in one window.

multiplot <- function(..., plotlist=NULL, file, cols=1, layout=NULL) {
  library(grid)

  # Make a list from the ... arguments and plotlist
  plots <- c(list(...), plotlist)

  numPlots = length(plots)

  # If layout is NULL, then use 'cols' to determine layout
  if (is.null(layout)) {
    # Make the panel
    # ncol: Number of columns of plots
    # nrow: Number of rows needed, calculated from # of cols
    layout <- matrix(seq(1, cols * ceiling(numPlots/cols)),
                    ncol = cols, nrow = ceiling(numPlots/cols))
  }

 if (numPlots==1) {
    print(plots[[1]])

  } else {
    # Set up the page
    grid.newpage()
    pushViewport(viewport(layout = grid.layout(nrow(layout), ncol(layout))))

    # Make each plot, in the correct location
    for (i in 1:numPlots) {
      # Get the i,j matrix positions of the regions that contain this subplot
      matchidx <- as.data.frame(which(layout == i, arr.ind = TRUE))

      print(plots[[i]], vp = viewport(layout.pos.row = matchidx$row,
                                      layout.pos.col = matchidx$col))
    }
  }
}

First, create plots. For example, p1=ggplot(), p2=ggplot(),…., p_n=ggplot(). The multiplot(p1,p2,...,p_n,cols= desires number of columns)
I ll demostrate multiplot in many examples

5 Barplot

5.1 One variable barplot

p1=ggplot(diamonds, aes(cut))+
 geom_bar(
   aes(y=(..count..)/sum(..count..)), # Percentage
          fill=c(brewer.pal(5,"Paired")))+
 coord_flip()+
 xlab("")+ylab("Percentage")+
 ggtitle("% of cut in diamonds dataset")

 # theme_wsj(base_size = 7.8,base_family = "Times", title_family = "Times")
 # +  scale_colour_wsj("colors6", "")
p2=ggplot(mtcars, 
          aes(gear))+
    geom_bar(
       aes(y=(..count..)), #Count
       fill=c(brewer.pal(3,"Paired")))+
    xlab("")+
    ylab("Count")+ggtitle("Count of gears in mtcars dataset")
multiplot(p1,p2,cols = 2)

Barplot example with diamonds and mtcar dataset. The first includes the percentage and the second the count

5.2 Two variable barplot: One numerical one factor

p3=ggplot(mtcars,
          aes(x=gear,y=disp,fill=as.factor(gear)))+
  geom_bar(stat="identity")+
  theme_bw()+
  guides(fill=FALSE) ## No theme and no legend

print(p3)

5.2.1 Marginal boxplot and multiplot

p4=ggMarginal(p = p3,#create a new plot based on an old one
           type = 'boxplot', #'boplot','histogram','density'
           margins = 'y',  # which axis i want marginal
           size = 5,  
           col = '#802E29', # colour of the lines
           fill = '#9C8219' 
)
print(p4)

Boxplot as marginal plot

How multiplot creates a marginal with both density and boxplot for disp variable

p5=ggMarginal(p = p3,#create a new plot based on an old one
           type = 'density', #'boplot','histogram','density'
           margins = 'y',  # which axis i want marginal
           size = 5,  
           col = '#802E29', # colour of the lines
           fill = '#9C8219' 
)
multiplot(p4,p5) # multiplot now creates a marginal with both density and boxplot

Now we created a marginal plot with both boxplot and barplot

5.3 Three variables barplot: One nymerical two factor and RcolourBrewer

Also introduce RColourBrewer package for colour palletes.This chart shows all the palletes included.

ggplot(mtcars,
       aes(x=as.factor(vs),y=disp,fill=as.factor(gear)))+
  geom_bar(stat="identity",
           position=position_dodge())+
  theme_bw()+
  scale_fill_brewer(palette="Dark2")

Three variable barplot

5.4 Legends, titles

In the above plots I did nothing to change the names of axis, legend or the limits of numerical variables. Lets use the barplot for three variables to see some examples.

p1=ggplot(mtcars,
           aes(x=as.factor(vs),
               y=disp,
               fill=as.factor(gear)))+
  geom_bar(stat="identity",
           position=position_dodge())+
  scale_fill_brewer(palette="Dark2",
                    name="Gears")+
  ggtitle("Three variable barplot with\n legends and labels plus limits")+
  ylab("disp variable")+
  xlab("vs")+
  scale_y_continuous(limits = c(0,600),
                     breaks = c(0,50,100,150,200,250,300,350,400,450,500,550,600))+
  theme(axis.text.y =element_text(angle = 45),
        panel.background = element_blank(),
        legend.position=c(0.96,0.85))

p2=ggplot(mtcars,
          aes(x=as.factor(vs),
              y=disp,
              fill=as.factor(gear)))+
  geom_bar(stat="identity",
           position=position_dodge())+
  scale_fill_brewer(palette="Accent",
                    name="Gears")+
  ggtitle("Three variable barplot with legends\n and labels plus limits")+
  ylab("disp variable")+
  xlab("vs")+
  scale_y_continuous(limits = c(0,600),
                     breaks = c(0,50,100,150,200,250,300,350,400,450,500,550,600))+
  theme(axis.text.y =element_text(angle = 45),
        panel.grid.major = element_blank(),
        legend.position=c(0.96,0.85),
        plot.title = element_text(hjust=0.5))



multiplot(p1,p2,cols = 2)

6 Boxplot, Violin plot and some themes

How fast and easy 3 plots can be created:

p1=ggplot(iris,aes(x=Species,y=Sepal.Length))+geom_violin()

p2=ggplot(iris,aes(x=Species,y=Sepal.Length))+geom_boxplot()

p3=ggplot(iris,aes(x=Species,y=Sepal.Length))+geom_tufteboxplot()

multiplot(p1,p2,p3,cols = 3)

Boxplot and violin plot for iris dataset

Lets add stuff to the violin and simple boxplot and start exploring themes as we go. Additional small changes like legend size, position, font colours, bold, title position and outliers/mean points

p1=ggplot(iris,
          aes(x=Species,
              y=Sepal.Length,
              fill=Species))+
  geom_violin()+
  ggtitle("Violin plot for iris data/economist theme")+
  theme_economist(base_size = 6)+
  theme(plot.title = element_text(hjust = 0.5),
        legend.position = "right",
        legend.background = element_rect( size=.3))



p2=ggplot(iris,
          aes(x=Species,
              y=Sepal.Length))+
  geom_boxplot()+
  theme_gdocs(base_size = 6)+
  ggtitle("Boxplot for iris data/gdocs theme")+
  theme(plot.title = element_text(hjust = 0.5))

p3=ggplot(iris,
          aes(x=Species,
              y=Sepal.Length,
              fill=Species))+
  geom_boxplot(outlier.colour="red",
               outlier.shape=3,
               outlier.size=2)+
  theme_wsj(base_size = 6)+
  ggtitle("wsj theme with outliers")+
  stat_summary(fun.y=mean,
               geom="point",
               shape=6,
               size=1)+
  scale_fill_brewer(palette="Dark2",
                    name="")+
  theme(plot.title = element_text(hjust = 0.5),
        legend.text = element_text(size=12,colour = "red"),
        legend.position = "bottom")+
  coord_flip() #+stat_compare_means(label.x=0.7,label.y =7,method = "t.test")#ggpubr package



p4=ggplot(iris,
          aes(x=Species,
              y=Sepal.Length,
              fill=Species))+
  geom_boxplot(outlier.colour="red",
               outlier.shape=3,
               outlier.size=2)+
  theme_dark(base_size = 8)+
  ggtitle("Dark theme")+
  stat_summary(fun.y=mean,
               geom="point",
               shape=6,
               size=1)+
  scale_fill_brewer(palette="Set1",
                    name="Species")+
  theme(plot.title = element_text(hjust = 0.5),  ## hjust centers the title of the plot
        legend.position=c(0.43,0.85),
        legend.text = element_text(colour="blue",
                                   size = 5,
                                   face = "bold"),
        legend.background = element_rect(fill="#7A7A7A",
                                         size=.5),
        legend.direction="horizontal")+
  xlab("")+
  ylab("") 

multiplot(p1,p2,p3,p4,cols = 2)

7 Scatter plots

Scatter plots are perfect for exploring relationships between two or more variables, numerical and factors.

7.1 Scatter matrix and ‘ggpairs’

This function I found on the internet is perfect for making scaterplot matrices

# another option as found on the internet
makePairs <- function(data) 
{
  grid <- expand.grid(x = 1:ncol(data), y = 1:ncol(data))
  grid <- subset(grid, x != y)
  all <- do.call("rbind", lapply(1:nrow(grid), function(i) {
    xcol <- grid[i, "x"]
    ycol <- grid[i, "y"]
    data.frame(xvar = names(data)[ycol], yvar = names(data)[xcol], 
               x = data[, xcol], y = data[, ycol], data)
  }))
  all$xvar <- factor(all$xvar, levels = names(data))
  all$yvar <- factor(all$yvar, levels = names(data))
  densities <- do.call("rbind", lapply(1:ncol(data), function(i) {
    data.frame(xvar = names(data)[i], yvar = names(data)[i], x = data[, i])
  }))
  list(all=all, densities=densities)
}
 
# expand iris data frame for pairs plot
gg1 = makePairs(iris[,-5])
 
# new data frame mega iris
mega_iris = data.frame(gg1$all, Species=rep(iris$Species, length=nrow(gg1$all)))
 
# pairs plot
ggplot(mega_iris, aes_string(x = "x", y = "y")) + 
  facet_grid(xvar ~ yvar, scales = "free") + 
  geom_point(aes(colour=Species), na.rm = TRUE, alpha=0.8) + 
  stat_density(aes(x = x, y = ..scaled.. * diff(range(x)) + min(x)), 
               data = gg1$densities, position = "identity", 
               colour = "grey20", geom = "line")

ggpairs() creates scatter plots for numerical, boxplot and barplots for factors and even includes correlations. Perfect for exploring small datasets fast. It is a little messy but lets see:

ggpairs(iris, title = "",  
  axisLabels = "show",
  aes(colour=Species),
  lower = list(combo = wrap(ggally_facethist,
                            binwidth = 0.4)))

- Some more ggpairs() options to make prettier or different plot. See Documentation:

ggpairs(iris,
        columns = c(1:4),
        title = "",
        aes(colour=Species),
        upper = list(continuous="cor",
                     binwidth=1),
        lower=list(continuous="smooth",
                   binwidth=1),
        switch = "y",
        axisLabels="none")+
  ggtitle("Only numerical values to create scattermatrix and corrplot with reggression lines(smooth) and 'few' theme")+
  theme_few(base_size = 8)

Only numerical variables

ggpairs(iris, 
        title = "",
        aes(colour=Species),
        upper = list(continuous="cor",
                     binwidth=0.3,
                     combo="facethist"),
        lower=list(continuous="smooth",
                   combo='box_no_facet',
                   binwidth=0.3),
        switch = "y",
        axisLabels="none")+
  ggtitle("Good Looking(?) Plot  Matrix with theme_calc and smaller letters for iris dataset")+
  theme_calc(base_size = 8)

All iris data variables

7.2 Scatterplot to examine relationship bettween vars

p1=ggplot(iris,aes(x=Sepal.Length,y=Petal.Length))+geom_point()

p2=p1+
    geom_smooth(method="lm")+
    ggtitle("Regression Line plus Standar Error")

p3=p1+
  geom_smooth(method="loess",
              linetype="dashed",
              se=F)+
  ggtitle("Loess dashed line without Standar error")+
  theme(plot.title =element_text(size=11))

pall=ggplot(iris,
            aes(x=Sepal.Length,
                y=Petal.Length))+
  geom_point(alpha = 0.3)+
  geom_smooth(method="lm")+
  geom_smooth(method="loess",
              linetype="dashed",
              se=F)+
  ggtitle("Loess dashed line without Standar error")+
  ggtitle("Simple Linear Regression line plus dashed loess line")+
  theme(plot.title =element_text(size=9))


p4=ggplot(mtcars,
          aes(x=mpg,
              y=hp,
              group=factor(cyl),
              color=factor(cyl)))+
  geom_point(size=1.5)+
  scale_colour_brewer(palette = "Set2",
                      name="cyl")+
  ggtitle("Three variables")+
  theme(plot.title=element_text(hjust=0.5))


p5=ggplot(mtcars,
          aes(x=mpg,
              y=hp,
              group=factor(cyl),
              color=factor(cyl)))+
  geom_point(size=2.5,
             aes(shape=factor(am)))+
  scale_colour_brewer(palette = "Set2",
                      name="cyl")+
  ggtitle("Four variables")+
  scale_shape(name="am")+
  ylab("")+
  theme(plot.title = element_text(hjust=0.5))

multiplot(p1,p2,p3,pall,cols = 2)

multiplot(p4,p5,cols=2)

8 More on themes (how to create custom themes)

You can create your own themes by modifying the existing complete themes. When you use a complete theme it overides the existing default characteristics. For example, theme_bw is a complete theme. If you use the theme option, as many times in previous plots, you can change specific plot elements .
So, to create a new custom theme just add stuff to a complete theme or replace all its elements. See the code example (press code).

# See the legend options of grey theme
theme_grey()$legend.key

## List of 5
##  $ fill         : chr "grey95"
##  $ colour       : chr "white"
##  $ size         : NULL
##  $ linetype     : NULL
##  $ inherit.blank: logi TRUE
##  - attr(*, "class")= chr [1:2] "element_rect" "element"

# Now lets change a specific element of the legend
new_theme <- theme_grey()+
             theme(legend.key = element_rect(colour = "red"))
new_theme$legend.key # It changes only the new defined element and keeps all the other characteristics the same

## List of 5
##  $ fill         : chr "grey95"
##  $ colour       : chr "red"
##  $ size         : NULL
##  $ linetype     : NULL
##  $ inherit.blank: logi FALSE
##  - attr(*, "class")= chr [1:2] "element_rect" "element"

# Now, with the %+replace% it replaces all options with the new options i define. So the undefined options remain NULL
new_theme <- theme_grey() %+replace% theme(legend.key = element_rect(colour = "red"))
new_theme$legend.key

## List of 5
##  $ fill         : NULL
##  $ colour       : chr "red"
##  $ size         : NULL
##  $ linetype     : NULL
##  $ inherit.blank: logi FALSE
##  - attr(*, "class")= chr [1:2] "element_rect" "element"

9 Interactivity

9.1 Plotly

Using plotly to make a ggplot reactive is too easy! It may change the appearence of the plot though.

require(plotly)
ggplotly(p4)

10 Interactive editing: R studio Addinns and ggplot2 User Interface

10.1 Package ‘ggplotgui’ - No code required!

It is a shiny app.

require(ggplotgui)

## Loading required package: ggplotgui

# Run this to open the UI
##   ggplot_shiny(mtcars)

GGplot2

Konstantinos

2017-09-16