Basic plots

Let’s define some sample data that we will work on:

sample_data <- data.frame(
  a = letters[1:10], 
  b = sample(x = 1:10, size = 10),
  color = sample(x = c("red", "green", "blue"), size = 10, replace = TRUE)
)

print(sample_data)

##    a  b color
## 1  a  7  blue
## 2  b  2   red
## 3  c  1 green
## 4  d  4 green
## 5  e  9 green
## 6  f  6   red
## 7  g  8 green
## 8  h  5  blue
## 9  i 10   red
## 10 j  3 green

The most basic plot:

library(ggplot2)
ggplot(data = sample_data, mapping = aes(x = a, y = b)) + 
    geom_point()

A little bit less basic plot, as points can be categorised by their colors:

ggplot(data = sample_data, mapping = aes(x = a, y = b, color = color)) +
    geom_point()

As you can see, colors do not match their descriptions, but you can customise it.

Here’s another way of separating categories:

ggplot(data = sample_data, mapping = aes(x = a, y = b)) + 
    geom_point() +
    facet_wrap(~ color, nrow=1)

Combining multiple types of plots

We’ll use a dataset mpg which is available in ggplot2 package.

ggplot(data = mpg) + 
    geom_smooth(mapping = aes(x = displ, y = hwy, color = drv, linetype=drv)) +
    geom_point(mapping = aes(x = displ, y = hwy, color = drv))

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

Smoothing may be useful if you want to show trend.

You can have different mapping for every plot:

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

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

Different datasets are also possible, but rather unusual.

Bar plots

The simplest bar plot:

ggplot(data = sample_data) +
    geom_bar(mapping = aes(x = a, y = b), stat = "identity")

We had po provide the argument stat = "identity", becasue the default behaviour is to plot the size/count of every category (x).

Stacked bar plot:

ggplot(data = sample_data, mapping = aes(x = color, y = b, color = a)) +
    geom_bar(stat = "identity", fill = NA)

Boxplot

Let’s use mpg data again:

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

We also used coord_flip(), which rotates the plot by 90 degrees, or, another words, flips the coordinates.

Maps

Let’s draw quickly a map of the USA:

usa <- map_data("usa")
ggplot(usa, aes(long, lat, group = group)) +
    geom_polygon(fill = "white", color = "black") +
    coord_quickmap()

We used two interesting functions:

• map_data() - a ggplot2’s function, which provides spatial data for a few countries in the world;

• coord_quickmap() - adjusts the size of a plot to the size of map. Default settings cause the opposite.

Other functions example

Here’s a weird plot, which aims at presenting various customisation examples:

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) +
    geom_rect(mapping=aes(xmin=15, xmax=20, ymin=0, ymax=max(hwy)), 
              fill='blue', alpha=0.1) +
    geom_point() + 
    labs(title = "Some plot",
         subtitle = "subtitle to chart",
         caption = "and caption: made by me",
         x = "city miles per gallon", 
         y = "highway miles per gallon") +
    geom_abline(color ="red") +  
    theme_bw() 

    theme(plot.title = element_text(hjust = 0.5, size=12),
          axis.title = element_text(size=12))

## List of 2
##  $ axis.title:List of 11
##   ..$ family       : NULL
##   ..$ face         : NULL
##   ..$ colour       : NULL
##   ..$ size         : num 12
##   ..$ hjust        : NULL
##   ..$ vjust        : NULL
##   ..$ angle        : NULL
##   ..$ lineheight   : NULL
##   ..$ margin       : NULL
##   ..$ debug        : NULL
##   ..$ inherit.blank: logi FALSE
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ plot.title:List of 11
##   ..$ family       : NULL
##   ..$ face         : NULL
##   ..$ colour       : NULL
##   ..$ size         : num 12
##   ..$ hjust        : num 0.5
##   ..$ vjust        : NULL
##   ..$ angle        : NULL
##   ..$ lineheight   : NULL
##   ..$ margin       : NULL
##   ..$ debug        : NULL
##   ..$ inherit.blank: logi FALSE
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  - attr(*, "class")= chr [1:2] "theme" "gg"
##  - attr(*, "complete")= logi FALSE
##  - attr(*, "validate")= logi TRUE

plotly

If you want to publish a plot on your website, consider using plotly:

p <- ggplot(sample_data, aes(x=a, y=b)) +
    geom_point()

plotly::ggplotly(p)

as it will give your plot interesting interactive features. Shiny users will appreciate them.

ssh

When you work on a remote machine and connect to it via ssh, the plots you create will not appear in pop-up windows by default. In order to do this, add the -X flag when connecting to server:

ssh -X user@login