Outline

• heatmap.2
• ggplot2
  • qplot()
  • ggplot()

ggplot2

ggplot2 is based on the grammer of graphics, the idea that you can build every graph from the same few components:

• data
• aesthetic mapping
• geometric object
• statistical transformations
• scales
• coordinate system
• position adjustments
• faceting

ggplot2 cheetsheet: https://www.rstudio.com/wp-content/uploads/2015/03/ggplot2-cheatsheet.pdf

ggplot2 grammers

ggplot() - graphics are added up by different layers

Compared to qplot(), it’s easier to use multiple dataset in ggplot().

• The basic usage is a combination of ggplot(), aes() and some geom function.
• ggplot() initializes a ggplot object. It can declare input data and a set of aesthetics.
• aes() describes how variables mapped to visual properties.
• geom function describe how the data will be visualized (point, line, histogram),
• Use “+” to add layers.
  • For example, ggplot(data) + aes(x = a, y = b) + geom point().

Aesthetics — aes()

• Colour related aesthetics: colour, fill and alpha
• Group related aesthetics: group
• Differentiation related aesthetics: linetype, size, shape
• Position related aesthetics: x, y, xmin, xmax, ymin, ymax, xend, yend

ggplot example

library(ggplot2)
ggplot(data = mpg) + aes(x=displ, y=hwy) + geom_point()

ggplot: combine layers

myggplot <- ggplot(data = mpg) + aes(x=displ, y=hwy)
myggplot + geom_point()

aes – color (continuous)

ggplot(data = mpg) + 
  aes(x=displ, y=hwy, color = cyl) +
  geom_point()

aes – color (categorical)

ggplot(data = mpg) + 
  aes(x=displ, y=hwy, color = class) +
  geom_point()

aes – size

ggplot(data = mpg) + 
  aes(x=displ, y=hwy, size = cyl) +
  geom_point()

aes – alpha (transparency)

ggplot(data = mpg) + 
  aes(x=displ, y=hwy, alpha = cyl) +
  geom_point()

aes – shape

mpg_sub <- subset(mpg, class!="suv") ## qplot support a maximum of 6 shapes
ggplot(data = mpg_sub) + 
  aes(x=displ, y=hwy, shape = class) +
  geom_point()

Geom functions

ls(pattern = '^geom_', env = as.environment('package:ggplot2'))

##  [1] "geom_abline"     "geom_area"       "geom_bar"       
##  [4] "geom_bin2d"      "geom_blank"      "geom_boxplot"   
##  [7] "geom_col"        "geom_contour"    "geom_count"     
## [10] "geom_crossbar"   "geom_curve"      "geom_density"   
## [13] "geom_density_2d" "geom_density2d"  "geom_dotplot"   
## [16] "geom_errorbar"   "geom_errorbarh"  "geom_freqpoly"  
## [19] "geom_hex"        "geom_histogram"  "geom_hline"     
## [22] "geom_jitter"     "geom_label"      "geom_line"      
## [25] "geom_linerange"  "geom_map"        "geom_path"      
## [28] "geom_point"      "geom_pointrange" "geom_polygon"   
## [31] "geom_qq"         "geom_qq_line"    "geom_quantile"  
## [34] "geom_raster"     "geom_rect"       "geom_ribbon"    
## [37] "geom_rug"        "geom_segment"    "geom_sf"        
## [40] "geom_smooth"     "geom_spoke"      "geom_step"      
## [43] "geom_text"       "geom_tile"       "geom_violin"    
## [46] "geom_vline"

ggplot: geom_line by group

ggplot(data = mpg) + 
  aes(displ, hwy, colour=class) + 
  geom_point(aes(size=cyl)) + 
  geom_line(aes(group = class))

smooth by group 1

ggplot(data = mpg) + 
  aes(displ, hwy) + 
  geom_point(aes(colour=class)) + 
  geom_smooth() 

## `geom_smooth()` using method = 'loess' and formula 'y ~ x'

smooth by group 2

ggplot(data = mpg) + 
  aes(displ, hwy) + 
  geom_point(aes(colour=class)) + 
  geom_smooth(method="lm") 

smooth by group 3

ggplot(data = mpg) + 
  aes(displ, hwy) + 
  geom_point(aes(colour=class)) + 
  geom_smooth(aes(group=class), method="lm") 

smooth by group 4

ggplot(data = mpg) + 
  aes(displ, hwy, colour = class) + ## lower level aes will be applied to all higher level aes
  geom_point() + 
  geom_smooth(method="lm") 

smooth by group 5

ggplot(data = mpg) + 
  aes(displ, hwy, colour = class) + ## lower level aes will be applied to all higher level aes
  geom_point() + 
  geom_smooth(method="lm", se = F, size = 2) 

ggplot() boxplot

mpgbox <- ggplot(data = mpg) + 
  aes(class, hwy) + 
  geom_boxplot(aes(fill=class))
mpgbox

ggplot() jitter

ggplot(data = mpg) + 
  aes(class, hwy, color=class) + 
  geom_jitter()

ggplot() boxplot + jitter

ggplot(data = mpg) + 
  aes(class, hwy, color=class) + 
  geom_boxplot() + 
  geom_jitter()

ggplot() violin plot

ggplot(data = mpg) + 
  aes(class, hwy, fill=class) + 
  geom_violin() 

ggplot() bar plot 1

ggplot(mpg) + 
  aes(class) + 
  geom_bar()

ggplot() bar plot 2

ggplot(mpg) + 
  aes(class, fill=as.factor(cyl)) + 
  geom_bar()

ggplot() bar plot 3

ggplot(mpg) + 
  aes(class, fill=as.factor(cyl)) + 
  geom_bar(position="dodge")  #side by side

ggplot() bar plot: how to specify error bar

mpgSummary <- data.frame(class = with(mpg, tapply(class, class, unique)), 
                         meanDispl = with(mpg, tapply(displ, class, mean)),
                         sdDispl = with(mpg, tapply(displ, class, sd)))

ggplot(data = mpgSummary) + 
  aes(x=class, y=meanDispl, fill=class) + 
  geom_bar(position=position_dodge(), stat="identity",
           colour="black", # Use black outlines,
           size=.3) +      # Thinner lines
  geom_errorbar(aes(ymin=meanDispl-sdDispl, ymax=meanDispl+sdDispl),
                size=.3,    # Thinner lines
                width=.2,
                position=position_dodge(.9))

ggplot() histogram simple example

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram()

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

ggplot() histogram fill by color

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram(aes(fill = class))

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

ggplot() histogram facets by group (1)

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram(aes(fill = class)) + 
  facet_wrap(~ class)

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

facet

• facet_grid: lay out panels in a grid.
• facet_wrap: wrap a 1d ribbon of panels into proper 2d display

ggplot() histogram facets by group (2)

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram(aes(fill = class)) + 
  facet_grid(. ~ class)

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

ggplot() histogram facets by group (3)

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram(aes(fill = class)) + 
  facet_grid(class ~ .)

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

ggplot() histogram facets by group (4)

ggplot(data = mpg) + 
  aes(x = hwy) + 
  geom_histogram(aes(fill = class)) + 
  facet_grid(drv ~ class)

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

longitudinal data visualization

• spaghetti plot
• individual trajectory
• mean trajectory (with se bar)

sleepstudy: Reaction times in a sleep deprivation study

• This laboratory experiment measured the effect of sleep deprivation on cognitive performance.
• There were 18 subjects, chosen from the population of interest (long-distance truck drivers), in the 10 day trial. These subjects were restricted to 3 hours sleep per night during the trial.
• On each day of the trial each subject’s reaction time was measured. The reaction time shown here is the average of several measurements.
• These data are balanced in that each subject is measured the same number of times and on the same occasions.

sleepstudy: Reaction times in a sleep deprivation study

library(lme4)

## Loading required package: Matrix

data(sleepstudy)
head(sleepstudy, n=5)

##   Reaction Days Subject
## 1 249.5600    0     308
## 2 258.7047    1     308
## 3 250.8006    2     308
## 4 321.4398    3     308
## 5 356.8519    4     308

spaghetti plot

ggplot(data=sleepstudy) + 
  aes(x = Days, y=Reaction, colour = Subject) +
  geom_path()

individual subject lm smooth

ggplot(data=sleepstudy) + 
  aes(x = Days, y=Reaction, colour = Subject) +
  geom_smooth(method="lm") + 
  facet_wrap(~Subject)

mean trajectory (with SD bar)

sleepSummary <- with(sleepstudy, 
                     data.frame(Days = tapply(Days,Days,unique),
                                Mean = tapply(Reaction,Days,mean),
                                SD = tapply(Reaction,Days,sd))
                     )

ggplot(data=sleepSummary) + 
  aes(x = Days, y=Mean) +
  geom_path() + 
  geom_errorbar(aes(ymin=Mean-SD, ymax=Mean+SD),
                  size=0.5,    # Thinner lines
                  width=.2) 

Add text annotations to a graph

• Ref: http://www.sthda.com/english/wiki/ggplot2-texts-add-text-annotations-to-a-graph-in-r-software
  • geom_text(): adds text directly to the plot
  • geom_label(): draws a rectangle underneath the text, making it easier to read.
  • annotate(): useful for adding small text annotations at a particular location on the plot
  • annotation_custom(): Adds static annotations that are the same in every panel

Create some data from mtchar

# Subset 10 rows
set.seed(32611)
ss <- sample(1:32, 10)
df <- mtcars[ss, ]

Text annotations using geom_text()

sp <- ggplot(data = df) +
  aes(wt, mpg, label = rownames(df)) +
  geom_point()
# Add texts
sp + geom_text()

Other experiment

• Change the size of the texts

sp + geom_text(size=6)

• Change vertical and horizontal adjustement

sp +  geom_text(hjust=0, vjust=0)

• Change fontface. Allowed values : 1(normal), 2(bold), 3(italic), 4(bold.italic)

sp + geom_text(aes(fontface=2))

• Change font family

sp + geom_text(family = "Times New Roman")

• Color by groups

sp + geom_text(aes(color=factor(cyl)))

• Set the size of the text using a continuous variable

sp + geom_text(aes(size=wt))

Text annotations using geom_label()

sp <- ggplot(data = df) +
  aes(wt, mpg, label = rownames(df)) +
  geom_point()
# Add texts
sp + geom_label()

Add a text annotation at a particular coordinate

# Solution 1
sp + geom_text(x=3, y=20, label="Scatter plot")

ggrepel: Avoid overlapping of text labels

• geom_label_repel()
• geom_text_repel()

library(ggrepel)

Create a scatter plot and add labels

p <- ggplot(mtcars, aes(wt, mpg)) +
  geom_point(color = 'red') 
p + geom_text(aes(label = rownames(mtcars)),
              size = 3.5)

Use geom_text_repel

set.seed(32611)
p + geom_text_repel(aes(label = rownames(mtcars)),
                    size = 3.5) 

Use label_text_repel

set.seed(32611)
p + geom_label_repel(aes(label = rownames(mtcars)))

## p + geom_label_repel(aes(label = rownames(mtcars), fill = factor(cyl)), 
##                         color = 'white', size = 3.5
##                    )

Labs

• title, x, y

p <- ggplot(mpg) + 
  geom_point(aes(x = displ, y = hwy, colour=factor(cyl))) + 
  labs(title = "New plot title", x = "New x label", y = "New y label")

Theme

• theme_grey: signature ggplot2 theme.
• theme_bw:dark-on-light theme.
• theme_linedraw: flip x and y.
• theme_light: light grey background.
• theme_dark: dark background.
• theme_minimal: no background annotations.
• theme_classic: classic-looking theme.
• theme_void: empty theme.

examples on different themes

p <- ggplot(mpg) + 
  geom_point(aes(x = displ, y = hwy, colour=factor(cyl))) + 
  facet_wrap(~class)
p ## default theme_grey

More about the theme

p + theme(text = element_text(size=20),
        axis.text.x = element_text(angle=90, hjust=1,colour="red")) 

Four elements to control the theme

• element_blank(): draws nothing, and assigns no space.
• element_rect(): borders and backgrounds.
• element_line(): lines
• element_text(): text

plot <- ggplot(mpg, aes(displ, hwy)) + geom_point()

plot + theme(
  panel.background = element_blank(),
  axis.text = element_blank()
)

plot + theme(
  axis.text = element_text(colour = "red", size = rel(1.5))
)

plot + theme(
  axis.line = element_line(arrow = arrow())
)

plot + theme(
  panel.background = element_rect(fill = "white"),
  plot.margin = margin(2, 2, 2, 2, "cm"),
  plot.background = element_rect(
    fill = "grey90",
    colour = "black",
    size = 1
  )
)
## all changes are relative to the default value

• Arguments in the theme function control figure theme
• There is a hierarchy in these arguments. (E.g., text will overwrite all text related arguments.)

line
rect
text
title
aspect.ratio
axis.title
axis.title.x
axis.title.y 
axis.text
axis.text.x
axis.text.y
axis.ticks
axis.ticks.x
axis.ticks.y,
axis.ticks.length
axis.line
axis.line.x
axis.line.y
## for more options, see
?theme

• look at theme_gray() function

theme_gray

No legend

ggplot(mpg) + 
  geom_point(aes(x = displ, y = hwy, colour=factor(cyl))) + 
  theme(legend.position = "none") 

One of my favourate themes (1)

black.bold.text <- element_text(face = "bold", color = "black", size=20)
ggplot(mpg, aes(displ, hwy, colour=class)) + geom_point() + 
    labs(title="hwy vs displ") + 
    theme_bw() + 
    theme(text = black.bold.text) 

One of my favourate themes (2)

black.bold.text <- element_text(face = "bold", color = "black", size=20)
ggplot(mpg, aes(displ, hwy, colour=class)) + geom_point() + 
    labs(title="hwy vs displ") + 
    theme_bw() + 
    theme(text = black.bold.text, panel.grid =element_blank()) 

Other experiments

black.bold.text <- element_text(face = "bold", color = "black", size=20)
red.italic.text <- element_text(face = "italic", color = "red", size=15)

ggplot(mpg, aes(displ, hwy, colour=class)) + geom_point() + 
    labs(title="hwy vs displ") + 
    theme_bw() + 
    theme(axis.text = black.bold.text , axis.title = black.bold.text, 
          legend.title = red.italic.text, 
          legend.text = black.bold.text) 

Stat transformation

• stat ecdf: Empirical Cumulative Density Function.
• stat ellipse: Plot data ellipses.
• stat function: Superimpose a function.
• stat identity: Identity statistic.
• stat qq: Calculation for quantile-quantile plot.
• stat summary 2d: Bin and summarise in 2d.
• stat unique: Remove duplicates.

empirical CDF

df <- data.frame(x = rnorm(1000))
ggplot(df, aes(x)) + stat_ecdf(geom = "step")

stat_function

n <- 100
set.seed(32611)
df <- data.frame(
  x = rnorm(n)
)
x <- df$x
base <- ggplot(df, aes(x)) + geom_density()
base + stat_function(fun = dnorm, colour = "red") + xlim(c(-3,3))

stat_ellipse

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() +   stat_ellipse()

stat_ellipse by group

ggplot(mpg, aes(x = displ, y = hwy, color=displ > 4)) + geom_point() +   stat_ellipse()

Coordinate

• coord_cartesian: Cartesian coordinates.
• coord_fixed: control_fixed ratio between x and y scales.
• coord_flip: flip x and y.
• coord_map: map projections.
• coord_polar: polar coordinates.
• coord_trans: transformed cartesian coordinate system.

Our old friend mpg

p <- ggplot(mpg, aes(displ, hwy)) +
  geom_point() +
  geom_smooth()

p

## `geom_smooth()` using method = 'loess' and formula 'y ~ x'

Setting the limits on the coordinate system performs a visual zoom.

p + coord_cartesian(xlim = c(3, 5), expand = FALSE)

## `geom_smooth()` using method = 'loess' and formula 'y ~ x'

Setting the limits on a scale converts all values outside the range to NA.

p + xlim(3, 5)

## `geom_smooth()` using method = 'loess' and formula 'y ~ x'

## Warning: Removed 136 rows containing non-finite values (stat_smooth).

## Warning: Removed 136 rows containing missing values (geom_point).

#the same as p + scale_x_continuous(limits = c(325, 500))

resize the plot

p <- ggplot(mpg, aes(displ, hwy)) +  geom_point()
p + coord_fixed(ratio = 0.5)

p + coord_fixed(ratio = 0.1)

flip x and y

ggplot(mpg, aes(class, hwy)) +
  geom_boxplot() +
  coord_flip()