Plotting

February 2019

1 / 60

2 / 60

`base` Plotting

The traditional approach for plotting.
Separate functons for different plots

# Histogram in base R
hist(x = basel$alter,
     xlab = "Alter",
     ylab = "Häufigkeit",
     main = "Histogramm Alter")

3 / 60

`base` Plotting

The traditional approach for plotting.
Separate functons for different plots

# Boxplot in base R
boxplot(formula = groesse ~ geschlecht,
        data = basel,
        xlab = "Geschlecht",
        ylab = "Groesse",
        main = "Box plot Groesse")

4 / 60

`base` Plotting

The traditional approach for plotting.
Separate functons for different plots

# Scatterplot in base R
plot(x = basel$groesse,
     y = basel$einkommen,
     xlab = "Height",
     ylab = "Einkommen",
     main = "Scatterplot Groesse x Einkommen")

5 / 60

Probleme mit Base R plotting

Aesthetics are outdated.
Pretty plots = lots of code.
Relatively inflexible.

from healthhosts.com

6 / 60

The mighty `tidyverse`

The tidyverse is a collection of high-performing, user-friendly R packages, created explicitly for efficient data analytics.

ggplot2 for graphics.
dplyr for data wrangling.
tidyr for data wrangling.
readr for data I/O.
purrr for function programming.
tibble for modern data.frames.

7 / 60

Modular graphics in ggplot2

data: The dataset
mapping: The structure of the plot

What should axes represent?
What should color/size represent?

geoms: Objects drawn within plot
labs: Plot annotation
themes: Aesthetic design
facets: Plot facets
scales: Scaling of axes

8 / 60

Part 1: Wir kreieren diesen Plot

data

The mpg dataset

mapping

Engine displacement onto x-axis
Miles per gallon onto y-axis
Color according to vehicle class

geoms

Plot data as points
Add regression line

labs

Annotation of axes and titles

themes

black-white aesthetics

9 / 60

`ggplot()`

All plots start with ggplot()
2 central arguments

data | The dataset (tibble)
mapping | The plot structure defined using aes()

ggplot(data = mpg)

10 / 60

`aes()`

aes() defines the structure for the mapping argument
Central arguments:

x,y | Defines axes
color,fill | Defines colors
alpha | Defines transparency
size | Defines size
shape | Defines object sbapes (e.g. circles or squares)

ggplot(data = mpg,
       mapping = aes(x = displ, y = hwy))

11 / 60

+

The + operator extents plots with additional modules.

ggplot(data = mpg,
       mapping = aes(x = displ,
                     y = hwy)) + 
  # Plotte Daten als Punkte
  geom_point()

12 / 60

`geom_*()`

geom_*() functions define the geometric objects used to represent the data.
Some common geoms:

geom_point() | for points
geom_bar() | for bars
geom_boxplot() | for box-plots
geom_count() | for points scaled according to frequency
geom_smooth() | for curves

13 / 60

`geom_count()`

ggplot(data = mpg,
  mapping = aes(x = displ, y = hwy)) + 
  geom_count()

`geom_bar()`

ggplot(data = mpg,
  mapping = aes(x = class)) + 
  geom_bar()

14 / 60

`geom_boxplot()`

ggplot(data = mpg,
  mapping = aes(x = class,
                y = hwy)) + 
  geom_boxplot()

`geom_violin()`

ggplot(data = mpg,
  mapping = aes(x = class,
                y = hwy)) + 
  geom_violin()

15 / 60

What's missing?

16 / 60

`aes()`

aes() defines the structure for the mapping argument
2 arguments:

x,y | Defines axes
color,fill | Defines colors

ggplot(data = mpg,
       mapping = aes(x = displ, 
                     y = hwy,
                     # Farbe gemäss Klasse
                     color = class)) +
  geom_point()

17 / 60

What's missing?

18 / 60

`geom_smooth()`

geom_smooth() fits curves to the data and adds them to the plot.
Central arguments:

method | Type of model fitted to data
color | Color

ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  # Add curve
  geom_smooth(col = "blue")

19 / 60

`geom_smooth()`

geom_smooth() fits curves to the data and adds them to the plot.
Central arguments:

method | Type of model fitted to data
color | Color

ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  # Add curve
  geom_smooth(col = "blue",
              method = "lm")

20 / 60

Inheritance

geoms inherit their settings from mapping.
Vererbte Settings can be overruled using the geoms own arguments.

ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  geom_smooth()

from catchingfire.ca

21 / 60

Inheritance

geoms inherit their settings from mapping.
Vererbte Settings can be overruled using the geoms own arguments.

ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  geom_smooth()

22 / 60

What's missing?

23 / 60

`labs()`

labs() defines all annotation of the plot.
Central arguments:

x,y | annotation of axes
title, subtitle | title and subtitle
caption | caption

ggplot(...) +
  labs(x = "Engine displacement in liters",
       y = "Highway miles per gallon",
       title = "MPG dataset",
       subtitle = "Cars with higher Eng...",
       caption = "Source: MPG dataset...")

24 / 60

What's missing?

25 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:

theme_gray()
theme_classic()
theme_void()
theme_excel() (ggthemes)
theme_economist() (ggthemes)
theme_bw()

ggplot(...) + theme_gray()

26 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:

theme_gray()
theme_classic()
theme_void()
theme_excel() (ggthemes)
theme_economist() (ggthemes)
theme_bw()

ggplot(...) + theme_classic()

27 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:

theme_gray()
theme_classic()
theme_void()
theme_excel() (ggthemes)
theme_economist() (ggthemes)
theme_bw()

ggplot(...) + theme_void()

28 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:

theme_gray()
theme_classic()
theme_void()
theme_excel() (ggthemes)
theme_economist() (ggthemes)
theme_bw()

ggplot(...) + theme_excel()

29 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:

theme_gray()
theme_classic()
theme_void()
theme_excel() (ggthemes)
theme_economist() (ggthemes)
theme_bw()

ggplot(...) + theme_economist()

30 / 60

Formatierung mit `theme_*()`

Mit theme() können alle Aspekte eines Plots mit einem vorgefertigten Satz an Einstellungen ästhetisch formatiert werden.
Einige themes:
- theme_gray()
- theme_classic()
- theme_void()
- theme_excel() (ggthemes)
- theme_economist() (ggthemes)
- theme_bw()

ggplot(...) + theme_bw()

31 / 60

Et voila!

ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy, 
                     col = class)) +
  geom_point() + 
  geom_smooth(col = "blue", 
              method = "lm")+
  labs(
    x = "Engine displacement in liters",
    y = "Highway miles per gallon",
    title = "MPG dataset",
    subtitle = "Cars with higher Eng...",
    caption = "Source: MPG dataset...") + 
  theme_bw()

32 / 60

Part 2: Going crazy

Create and alter a gg -object.
Split plot in facets
Use themes to alter all aesthetics aspects of your plot
Create your own themes.
Save your plot as .pdf or .png.

from lonniemillsap.com

33 / 60

The gg object

The ggplot function generates a gg object that can be assigned
gg objects can be altered or extended using +.
Running the object generates the plot.

# Assign plot
my_plot <- ggplot(data = mpg,
                    aes(x = displ,y = hwy)) + 
  geom_point() + theme_bw()
# Show class
class(my_plot)

[1] "gg"     "ggplot"

my_plot

34 / 60

The gg object

The ggplot function generates a gg object that can be assigned
gg objects can be altered or extended using +.
Running the object generates the plot.

# Assign plot
my_plot <- ggplot(data = mpg,
                    aes(x = displ,y = hwy)) + 
  geom_point() + theme_bw()
# show class
class(my_plot)

[1] "gg"     "ggplot"

my_plot + geom_smooth()

35 / 60

`facet_*()`

Facetting creates the same plot for groups defined by a third variable.
Facet functions:

facet_wrap()
facet_grid()

# Without facetting
ggplot(data = mpg, 
       mapping = aes(x = displ, 
                     y = hwy)) + 
  geom_point() + theme_bw()

36 / 60

`facet_wrap()`

Facetting creates the same plot for groups defined by a third variable.
Facet functions:

facet_wrap()
facet_grid()

# Without facetting
ggplot(data = mpg, 
       mapping = aes(x = displ, 
                     y = hwy)) + 
  geom_point() + theme_bw() +
  facet_wrap(~ class)

37 / 60

`facet_grid()`

Facetting creates the same plot for groups defined by a third variable.
Facet functions:

facet_wrap()
facet_grid()

# Without facetting
ggplot(data = mpg, 
       mapping = aes(x = displ, 
                     y = hwy)) + 
  geom_point() + theme_bw() +
  facet_grid(drv ~ class)

38 / 60

`theme()`

87 arguments in theme() can be used to alter all aesthetic aspects of a plot.
Requires helper functions:

element_rect() | for areas
element_line() | for lines
element_text() | for text
element_blank() | for removals

# Use theme
my_plot + 
  theme(argument = element_*(), 
        argument = element_*(),
        ...)

39 / 60

Background

Arguments for the background:

panel.background | for the inner background
plot.background | for the outer background

# Change background
my_plot + 
  theme(
    panel.background = 
      element_rect(fill = 'tomato'), 
    plot.background = 
      element_rect(fill = 'burlywood'))

40 / 60

Grid

Arbguments for the grid:

panel.grid.major | larger grid lines
panel.grid.minor | smaller grid lines

# Change grid
my_plot + 
  theme(
    panel.grid.major = 
      element_line(colour = "salmon"),
    panel.grid.minor = 
      element_line(colour = "seagreen"))

41 / 60

Raster

Arbguments for the grid:

panel.grid.major | larger grid lines
panel.grid.minor | smaller grid lines

# Change grid
my_plot + 
  theme(
    panel.grid.major = 
      element_line(colour = "salmon",
                   size = 3),
    panel.grid.minor = 
      element_line(colour = "seagreen",
                   size = 1.5))

42 / 60

Axes

Arguments for axes:

axis.line.x | x-axis
axis.line.y | y-axis
axis.title.x | x-axis title
axis.title.y | y-axis title

# Change axes
my_plot + 
  theme(
    axis.line.x = 
      element_line(colour = "deeppink", 
                   size = 3.5, 
                   lineend = "butt"),
    axis.line.y = 
      element_line(colour = "deeppink", 
                   size = 3.5))

43 / 60

Axes

Arguments for axes:

axis.line.x | x-axis
axis.line.y | y-axis
axis.title.x | x-axis title
axis.title.y | y-axis title

# Change axes titles
my_plot + 
  theme(
    axis.title.x = 
      element_text(family = "Comic Sans MS", 
                   size = 30),
    axis.title.y = 
      element_text(family = "Comic Sans MS", 
                   size = 30))

44 / 60

Further arguments (incomplete)

theme()

Argument	Description
axis.title.*	Everything concerning axes titles
axis.ticks.*	Everything concerning axes tick marks
axis.line.*	Everything concerning axis lines
legend.*	Everything concerning legends
panel.*	Everything concerning the inner plot region
plot.*	Everything concerning the outer plot region
strip.*	Everything concerning the facet headers

element_rect()

Argument	Description
fill	Color for filling areas
colour	Color for borders

element_line()

Argument	Description
size	Line sizes
linetype	Type of line

element_text()

Argument	Description
face	Font type (e.g., italic or bold)
colour	Font color

45 / 60

Own themes

my_theme <- theme(
  panel.background = 
    element_rect(fill = 'tomato'), 
  plot.background = 
    element_rect(fill = 'burlywood'),
  panel.grid.major = element_line(
    colour = "salmon", size = 3),
  panel.grid.minor = element_line(
    colour = "seagreen", size = 1.5),
  axis.line.x = element_line(
    colour = "deeppink", size = 3.5, 
    lineend = "butt"),
  axis.line.y = element_line(
    colour = "deeppink", size = 3.5),
  axis.title.x = element_text(
    family = "Comic Sans MS", size = 30),
  axis.title.y = element_text(
    family = "Comic Sans MS", size = 30))

my_plot

46 / 60

Own themes

my_theme <- theme(
  panel.background = 
    element_rect(fill = 'tomato'), 
  plot.background = 
    element_rect(fill = 'burlywood'),
  panel.grid.major = element_line(
    colour = "salmon", size = 3),
  panel.grid.minor = element_line(
    colour = "seagreen", size = 1.5),
  axis.line.x = element_line(
    colour = "deeppink", size = 3.5, 
    lineend = "butt"),
  axis.line.y = element_line(
    colour = "deeppink", size = 3.5),
  axis.title.x = element_text(
    family = "Comic Sans MS", size = 30),
  axis.title.y = element_text(
    family = "Comic Sans MS", size = 30))

my_plot + my_theme

47 / 60

`scale_*()`

scale_*() funcgions adjust all sizes and lengths.
Types of scale_*() functions:

scale_xy_* | Scale axes
scale_color_*, scale_color_* | Scale colors
scale_size_* | Scale sizes
scale_alpha_* | Scale transparency
...

my_plot

48 / 60

`scale_x_continuous()`

scale_*() funcgions adjust all sizes and lengths.
Types of scale_*() functions:

scale_xy_* | Scale axes
scale_color_*, scale_color_* | Scale colors
scale_size_* | Scale sizes
scale_alpha_* | Scale transparency
...

my_plot + 
  scale_x_continuous(limits = c(1, 30))

49 / 60

`scale_x_reverse()`

scale_*() funcgions adjust all sizes and lengths.
Types of scale_*() functions:

scale_xy_* | Scale axes
scale_color_*, scale_color_* | Scale colors
scale_size_* | Scale sizes
scale_alpha_* | Scale transparency
...

my_plot + 
  scale_x_reverse()

50 / 60

`scale_color_hue()`

scale_*() funcgions adjust all sizes and lengths.
Types of scale_*() functions:

scale_xy_* | Scale axes
scale_color_*, scale_color_* | Scale colors
scale_size_* | Scale sizes
scale_alpha_* | Scale transparency
...

my_plot + 
  scale_colour_hue(h = c(160, 260))

51 / 60

`scale_size()`

scale_*() funcgions adjust all sizes and lengths.
Types of scale_*() functions:

scale_xy_* | Scale axes
scale_color_*, scale_color_* | Scale colors
scale_size_* | Scale sizes
scale_alpha_* | Scale transparency
...

my_plot + 
  scale_size(range = c(1, 15))

52 / 60

Multiple plots

The patchwork package delivers a novel Syntax to join multiple plots in a single one.
patchwork syntax:

+ | join plots
| | join horizontally
/ | join vertically
() | bind plots
& | apply to all

# Speichere plots
pretty <- my_plot
ugly <- my_plot + my_theme

pretty + ugly

53 / 60

Multiple plots

The patchwork package delivers a novel Syntax to join multiple plots in a single one.
patchwork syntax:

+ | join plots
| | join horizontally
/ | join vertically
() | bind plots
& | apply to all

# Speichere plots
pretty <- my_plot
ugly <- my_plot + my_theme

pretty | ugly + pretty

54 / 60

Multiple plots

The patchwork package delivers a novel Syntax to join multiple plots in a single one.
patchwork syntax:

+ | join plots
| | join horizontally
/ | join vertically
() | bind plots
& | apply to all

# Speichere plots
pretty <- my_plot
ugly <- my_plot + my_theme

(pretty+pretty) / (pretty+pretty)

55 / 60

Multiple plots

The patchwork package delivers a novel Syntax to join multiple plots in a single one.
patchwork syntax:

+ | join plots
| | join horizontally
/ | join vertically
() | bind plots
& | apply to all

# Speichere plots
pretty <- my_plot
ugly <- my_plot + my_theme

(pretty+pretty) / (pretty+pretty) &
  my_theme

56 / 60

Multiple plots

The patchwork package delivers a novel Syntax to join multiple plots in a single one.
patchwork syntax:

+ | join plots
| | join horizontally
/ | join vertically
() | bind plots
& | apply to all

# Speichere plots
pretty <- my_plot
ugly <- my_plot + my_theme

(pretty+pretty) / (pretty+pretty) +
  plot_annotation(tag_levels = "A") & 
  theme(legend.position = "none")

57 / 60

  
          cdsbasel.github.io/dataanalytics/
        
       Data Analytics for Psychology and Business | February 2019
      
ggsave()Create graphic files with ggsave().
ggsave arguments:
filename | name of file
device | z.B. ".pdf" or ".png"
path | path to folder
height, width | height, width
unit | Unit for height, width
dpi | Resolution per unit

# Create plot
my_plot <- ggplot(data = mpg,
                 aes(x = displ, y = hwy)) + 
            geom_point() + 
            mytheme
# Save "my_plot.pdf"
ggsave(filename = "my_plot",
       plot = my_plot,
       device = "pdf",
       path = "figures",
       width = 6, 
       height = 4)

58 / 60

  
          cdsbasel.github.io/dataanalytics/
        
       Data Analytics for Psychology and Business | February 2019
      
ggsave()Create graphic files with ggsave().
ggsave arguments:
filename | name of file
device | z.B. ".pdf" or ".png"
path | path to folder
height, width | height, width
unit | Unit for height, width
dpi | Resolution per unit

# Create plot
my_plot <- ggplot(data = mpg,
                 aes(x = displ, y = hwy)) + 
            geom_point() + 
            mytheme
# Save "my_plot.pdf"
ggsave(filename = "my_plot",
       plot = my_plot,
       device = "png",
       path = "figures",
       width = 6, 
       height = 4)

59 / 60

Practical

60 / 60

Help

Keyboard shortcuts

↑, ←, Pg Up, k

Go to previous slide

↓, →, Pg Dn, Space, j

Go to next slide

Home

Go to first slide

End

Go to last slide

Number + Return

Go to specific slide

b / m / f

Toggle blackout / mirrored / fullscreen mode

Clone slideshow

Toggle presenter mode

Restart the presentation timer

?, h

Toggle this help

Plotting

Data Analytics for Psychology and Business

February 2019

base Plotting

base Plotting

base Plotting

Probleme mit Base R plotting

The mighty tidyverse

Modular graphics in ggplot2

Part 1: Wir kreieren diesen Plot

ggplot()

aes()

+

geom_*()

geom_count()

geom_bar()

geom_boxplot()

geom_violin()

What's missing?

aes()

What's missing?

geom_smooth()

geom_smooth()

Inheritance

Inheritance

What's missing?

labs()

What's missing?

Formatierung mit theme_*()

Formatierung mit theme_*()

Formatierung mit theme_*()

Formatierung mit theme_*()

Formatierung mit theme_*()

Formatierung mit theme_*()

Et voila!

Part 2: Going crazy

The gg object

The gg object

facet_*()

facet_wrap()

facet_grid()

theme()

Background

Grid

Raster

Axes

Axes

Further arguments (incomplete)

Own themes

Own themes

scale_*()

scale_x_continuous()

scale_x_reverse()

scale_color_hue()

scale_size()

Multiple plots

Multiple plots

Multiple plots

Multiple plots

Multiple plots

ggsave()

ggsave()

Practical

Help

Data Analytics for Psychology and Business
   

`base` Plotting

`base` Plotting

`base` Plotting

The mighty `tidyverse`

`ggplot()`

`aes()`

`geom_*()`

`geom_count()`

`geom_bar()`

`geom_boxplot()`

`geom_violin()`

`aes()`

`geom_smooth()`

`geom_smooth()`

`labs()`

Formatierung mit `theme_*()`

Formatierung mit `theme_*()`

Formatierung mit `theme_*()`

Formatierung mit `theme_*()`

Formatierung mit `theme_*()`

Formatierung mit `theme_*()`

`facet_*()`

`facet_wrap()`

`facet_grid()`

`theme()`

`scale_*()`

`scale_x_continuous()`

`scale_x_reverse()`

`scale_color_hue()`

`scale_size()`

`ggsave()`

`ggsave()`