Basics

library(dplyr)
library(ggplot2)
library(ggthemes)

The ggplot2 package (Wickham and Chang 2016) is based on “The Grammar of Graphics” (Wilkinson 2005). This theoretical framework helps us to construct statistical graphics by specifying several components. See https://bookdown.org/fjmcgrade/ismaykim/3-viz.html

data(iris)
glimpse(iris)

## Observations: 150
## Variables: 5
## $ Sepal.Length <dbl> 5.1, 4.9, 4.7, 4.6, 5.0, 5.4, 4.6, 5.0, 4.4, 4.9,...
## $ Sepal.Width  <dbl> 3.5, 3.0, 3.2, 3.1, 3.6, 3.9, 3.4, 3.4, 2.9, 3.1,...
## $ Petal.Length <dbl> 1.4, 1.4, 1.3, 1.5, 1.4, 1.7, 1.4, 1.5, 1.4, 1.5,...
## $ Petal.Width  <dbl> 0.2, 0.2, 0.2, 0.2, 0.2, 0.4, 0.3, 0.2, 0.2, 0.1,...
## $ Species      <fct> setosa, setosa, setosa, setosa, setosa, setosa, s...

ggplot(data=iris, aes(x = Sepal.Length, y = Sepal.Width)) + 
  geom_point(aes(color=Species)) +
  xlab("Sepal Length") +  
  ylab("Sepal Width") +
  ggtitle("Sepal Length-Width")

Let’s view this plot through the grammar of graphics:

The data variable Sepal Length gets mapped to the x-position aesthetic of the points.
The data variable Sepal Width gets mapped to the y-position aesthetic of the points.
The data variable Species gets mapped to the color aesthetic of the points.

The data variables correspond to columns in the iris data frame.
Note that data has to be in data frame format!
The geometric object considered here is of type point, but there are other types like lines, bars...

Data variable	Aes	Geom
Sepal Length	x	point
Sepal Width	y	point
Species	color	point

How to construct a plot?

Specify your data

ggplot(data=iris)

Specify mapping of variables to aesthetic components

ggplot(data=iris, mapping= aes(x=Sepal.Length, y= Sepal.Width))

Add layer : specify geometric object type

ggplot(data=iris, aes(x=Sepal.Length, y= Sepal.Width)) +
  geom_point()

Add other layers like title, labels and theme

ggplot(data=iris, aes(x=Sepal.Length, y= Sepal.Width)) +
  geom_point() +
  ggtitle("Sepal Length-Width")

Exercise 1 : warm up

Checks:

Title
Labels
Shape
Color
Transparency

Take away :

Note that there is no aes() surrounding alpha = 0.5 and color="red" here. Since we are NOT mapping a variable to an aesthetic but instead are just changing a setting, we don’t need to create a mapping with aes().
To improve legibility of your code, it's recommended to start a new line whenever adding a layer.
Note that your have to put the + sign always at the end of your line.

Update data layers

Create your base plot and save it as a variable.

p1<- ggplot(data=iris, aes(x = Sepal.Length, y = Sepal.Width)) + 
  geom_point(aes(color=Species)) +
  xlab("Sepal Length") +  
  ylab("Sepal Width") +
  ggtitle("Sepal Length-Width")

Use %+% to update your data layer.

p2 <- p1 %+% aes(shape=Species)
p2

p2 %+%
  aes(y=Petal.Length) %+% 
  ggtitle(" ")

Add a data layer.

p3 <- p2 + geom_smooth(method="lm", se=F)
p3

p4 <- p3 + facet_wrap(~Species)
p4

It's useful to also save commonly used data layers and combine them easily.

fw <- facet_wrap(~Species)
p1 + fw

p2 + fw

Facets

library(nycflights13)
mia_flights <- flights %>% 
  filter(dest == "MIA", !is.na(arr_delay))

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

We start with a basic histogram.

h1<- ggplot(data=mia_flights,aes(x=air_time)) +
       geom_histogram()  
h1

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

We split it into groups.

h1 +  facet_wrap(~origin)

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

You can play with the scale.

fw<- facet_wrap(~origin,scales='free_x')
h3<- h1 + fw
h3

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Update a layer.
And play with facet_wrap & facet_grid.

h4<-h3  %+% aes(fill=origin)
h4

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

h4 + facet_grid(carrier~origin,scales='free_x')

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

h4 + facet_wrap(carrier~origin, scales='free_x')

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Take away :

What is the difference between facet_grid & facet_wrap?

Facet_grid(x~y) will show all **x*y plots** even if a plot is empty.
Facet_wrap(x~y) only shows plots having actual values.

Scales can be made independent, by setting them 'free'.

Exercise 2: How does air_time evolves?

Highlight the chick!

We can plot the evolution of the weight of chicks over time.

data(ChickWeight)
ggplot() +
  geom_line(data=ChickWeight, aes(x=Time, y=weight, group=Chick), color = "gray")

But tell me, how does the weight of Chick 17 evolve?

ggplot() +
  geom_line(data=ChickWeight, aes(x=Time, y=weight, group=Chick), color = "gray") +
  geom_line(data=subset(ChickWeight, Chick==17),
            aes(x=Time, y=weight, group=Chick), color = "red", size = 1) +
  labs(title = "Weight of Chicks 17 versus other chicks")

Select multiple chicks upfront to highlight them.

selected_chicks <- ChickWeight %>%
  filter(Chick %in% c(15, 16, 17))

ggplot(data=ChickWeight, aes(x=Time, y=weight, group=Chick)) +
  geom_line(color = "gray") +
  geom_line(data=selected_chicks, aes(color = Chick), size = 1) +
  labs(title = "Weight of Chicks 15, 16, 17 versus other chicks")

Themes

Let's create the ugliest plot you've ever seen

In the theme layer you specify all visual elements that are not part of the data like text, lines and rectangels.

Let's go back to our original ChickWeight plot

cw<-ggplot() +
  geom_line(data=ChickWeight, aes(x=Time, y=weight, group=Chick), color = "gray") 

cw

And let's change the background color to green.

cw + theme(plot.background=element_rect(fill="green"))

Add a thick red border.

cw + theme(plot.background=element_rect(fill="green", color="red", size=5))

Let's get rid of the grey panel background.

cw + theme(plot.background=element_rect(fill="green", color="red", size=5),
           panel.background=element_blank())

You will love element_blank()! Do you really need those gridlines?

cw + theme(plot.background=element_rect(fill="green", color="red", size=5),
           panel.background=element_blank(),
           panel.grid = element_blank())

Let's make this plot even more ugly by adding blue lines and yellow ticks.

cw + theme(plot.background=element_rect(fill="green", color="red", size=5),
           panel.background=element_blank(),
           panel.grid = element_blank(),
           axis.line=element_line(color="blue"),
           axis.ticks=element_line(color="yellow"))

To finish, we should definitely pimp the x axis label!

cw + theme(plot.background=element_rect(fill="green", color="red", size=5),
           panel.background=element_blank(),
           panel.grid = element_blank(),
           axis.line=element_line(color="blue"),
           axis.ticks=element_line(color="yellow"),
           axis.title.x=element_text(color="darkblue",hjust=0,face="italic"))

We should definitely save our theme, because we want to apply it to all our plots!

ugly_theme <- theme(plot.background=element_rect(fill="green", color="red", size=5),
                    panel.background=element_blank(),
                    panel.grid = element_blank(),
                    axis.line=element_line(color="blue"),
                    axis.ticks=element_line(color="yellow"),
                    axis.title.x=element_text(color="darkblue",hjust=0,face="italic"))

Remember our 'iris plot'?

iris_plot <- ggplot(data=iris, aes(x=Sepal.Length, y= Sepal.Width)) +
  geom_point(aes(shape=Species), alpha=0.5, color="red") +
  xlab("Sepal Length") +
  ylab("Sepal Width") +
  ggtitle("Sepal Length-Width")

iris_plot

iris_plot + ugly_theme

If you want to change some other things not included in your theme, you can just add another theme_layer.

iris_plot + ugly_theme + theme(legend.background = element_blank(),
                               legend.key  = element_blank())

Exercise 3

Use the theme_layer to convert our original iris_plot to the following one.

Ggthemes

http://www.ggplot2-exts.org/ggthemes.html
https://github.com/jrnold/ggthemes

iris_plot + theme_bw()

Note that you can just add your own theme preferences upon an existing theme.
Here we use theme_bw(), but we want to remove the ticks.

iris_plot + theme_bw() + theme(axis.ticks=element_blank())

To know what's in a theme.

theme_bw

## function (base_size = 11, base_family = "", base_line_size = base_size/22, 
##     base_rect_size = base_size/22) 
## {
##     theme_grey(base_size = base_size, base_family = base_family, 
##         base_line_size = base_line_size, base_rect_size = base_rect_size) %+replace% 
##         theme(panel.background = element_rect(fill = "white", 
##             colour = NA), panel.border = element_rect(fill = NA, 
##             colour = "grey20"), panel.grid = element_line(colour = "grey92"), 
##             panel.grid.minor = element_line(size = rel(0.5)), 
##             strip.background = element_rect(fill = "grey85", 
##                 colour = "grey20"), legend.key = element_rect(fill = "white", 
##                 colour = NA), complete = TRUE)
## }
## <environment: namespace:ggplot2>

Did you know that there's also an RLadies theme?
You can find it on Github

https://github.com/rladies/starter-kit/blob/master/rladiesggplot2theme.R

Let's make our 'iris plot' RLadies-style!

iris_plot + r_ladies_theme()

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

## Warning in grid.Call.graphics(C_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database

## Warning in grid.Call.graphics(C_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database

## Warning in grid.Call.graphics(C_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database

Zoom on Barplots (by Huong)

The default barplot lay-out is verticle.

ggplot(pct.crash.table, aes(x = BOROUGH, y = total.injured)) +
  # make bar graphs; default is verticle
  geom_bar(stat = "identity")

You can change it to horizontal by using coord_flip().

ggplot(pct.crash.table, aes(x = BOROUGH, y = total.injured)) +
  # change the colors of bars
  geom_bar(stat = "identity", colour = "orange") +
  # horizontal bars
  coord_flip()

You can set colour: it will set the outline colour of your bars.
You can set fill : it will set the colour of your bars.

ggplot(pct.crash.table, aes(x = BOROUGH, y = total.injured)) +
  # change the colors of bars
  geom_bar(stat = "identity", colour = "orange", fill="orange") +
  # horizontal bars
  coord_flip()

You can reorder your data in the data layer.
And change the labels of x and y axis.

ggplot(pct.crash.table, aes(x = reorder(BOROUGH, total.injured), y = total.injured)) +
  geom_bar(stat = "identity", colour = "orange", fill ="orange") +
  coord_flip() +
  labs(x = "borough",
       y = "total injured (%)")

How do we create following plot?

From this plot

To this plot

Let's go step by step!

Set the data layer.

m1<- ggplot(modes.injured2, aes(x = reorder(borough1, -value2), y = value, fill = variable))
m1

Add geom_bar. Note that we set the colour.

m1 + geom_bar(stat = "identity", colour = "gray100")

We want to change the width.

m1 + geom_bar(stat = "identity", colour = "gray100", width = .6)

And change the position.

m2 <- m1 + geom_bar(stat = "identity", colour = "gray100", width = .6, position = "dodge")
m2

Let's flip the axes.

m3 <- m2 + coord_flip()
m3

Let's add some data labels!

m4 <- m3 +  geom_text(aes(label = paste0(round(modes.injured2$value,0), "%")),
                      position = position_dodge(width = 0.5),
                      hjust = - .2, 
                      size = 3)
m4

Let's manually define the colors.

modes.colors = c("motorists" = "sienna",
                 "pedestrians" = "sienna1",
                 "cyclists" = "sienna3")

m5 <- m4 + scale_fill_manual("", values = modes.colors)
m5

Enlarge the y-axis manually from 0 to 50.

m6 <- m5 + scale_y_continuous(limits = c(0,50)) 
m6

Move the legend to the bottom.

m7 <- m6 + 
  theme(legend.position = "bottom",
        legend.text = element_text(size = 10))
m7

Get rid of the grey background and add lines for 0, 10, 20..

m8 <- m7 +
  guides(fill = guide_legend(nrow = 1, byrow = TRUE, reverse = FALSE)) +
  theme(axis.text.x =element_text(size = 7),
        axis.title.x = element_blank(),
        axis.line.x = element_blank(),
        axis.text.y =element_text(size = 12),
        axis.title.y = element_blank(),
        axis.ticks.y = element_blank(),
        axis.line.y = element_blank(),
        panel.grid.major.x = element_line(colour = "azure3", size = 0.2),
        panel.background = element_blank(),
        panel.spacing = unit(4, "lines"))
m8

Putting graphs side by side.

m9<- m8 + facet_grid(. ~ variable5, scales = "fixed", space = "free")
m9

Change facet label.

m10 <- m9 + theme(strip.background = element_blank(),
                  # put facet label to the left of axis label
                  strip.placement = "outside",
                  strip.text.x = element_text(angle = 0,colour= "black",size = unit(14,'pt')))
m10

Extra

Heatmap

ggplot(data=me,aes(x=hour,y=wd)) + 
  geom_tile(aes(fill=avg_dist)) + 
  scale_fill_gradientn(colors=c( "darkblue", "orange","yellow"), name="minutes") + 
  labs(title="Average delay in arrival when flying from JFK airport in 2013",
       subtitle="By hour and weekday",
       caption = "When would you book your flight?") +
  theme(
    panel.background=element_blank(), 
    axis.ticks=element_blank(),
    axis.title=element_blank()
  )

Extensions

There are some great ggplot extensions http://www.ggplot2-exts.org/gallery/

Add labels to your points with ggrepel

ggplot(data=mia_flights, aes(x=dep_delay, y= arr_delay, color=carrier)) +
  geom_point() +
  geom_text(data=filter(mia_flights, dep_delay>500), aes(label=origin))

You can avoid overlapping by using the ggrepel package.

library(ggrepel)
ggplot(data=mia_flights, aes(x=dep_delay, y= arr_delay, color=carrier)) +
  geom_point() +
  geom_text_repel(data=filter(mia_flights, dep_delay>500), aes(label=origin))

GGally

How is my data distributed?

library(GGally)
ggpairs(iris)

Patchwork

Combine plots how you want!
https://github.com/thomasp85/patchwork

library(patchwork)
p1<-ggplot(data=iris, aes(x = Sepal.Length, y = Sepal.Width)) + 
  geom_point(aes(color=Species)) +
  xlab("Sepal Length") +  
  ylab("Sepal Width") +
  ggtitle("Sepal Length-Width")
p2 <- p1 %+% aes(shape=Species) + geom_smooth(method="lm", se=F)
p1 + p2

Colors

Do you need some ideas for great matching colors?
https://colorbrewer2.org

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Ggplot_tutorial.md

Ggplot_tutorial.md

Basics

How to construct a plot?

Exercise 1 : warm up

Update data layers

Facets

Exercise 2: How does air_time evolves?

Highlight the chick!

Themes

Let's create the ugliest plot you've ever seen

Exercise 3

Ggthemes

Zoom on Barplots (by Huong)

How do we create following plot?

Extra

Heatmap

Extensions

Add labels to your points with ggrepel

GGally

Patchwork

Colors

Files

Ggplot_tutorial.md

Latest commit

History

Ggplot_tutorial.md

File metadata and controls

Basics

How to construct a plot?

Exercise 1 : warm up

Update data layers

Facets

Exercise 2: How does air_time evolves?

Highlight the chick!

Themes

Let's create the ugliest plot you've ever seen

Exercise 3

Ggthemes

Zoom on Barplots (by Huong)

How do we create following plot?

Extra

Heatmap

Extensions

Add labels to your points with ggrepel

GGally

Patchwork

Colors