ggplots

Using tidyverse for ggplots

1. load tidyverse

note: for some reason you have to install tidyverse on Rcloud every time, so here’s the code for easier access:

install.packages("tidyverse")

## Installing package into '/cloud/lib/x86_64-pc-linux-gnu-library/4.0'
## (as 'lib' is unspecified)

• using ctrl+L is useful to erase all the ugly code that comes from that

• We can now use tidyverse!

library(tidyverse)

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.4     ✔ readr     2.1.5
## ✔ forcats   1.0.0     ✔ stringr   1.5.1
## ✔ ggplot2   3.5.1     ✔ tibble    3.2.1
## ✔ lubridate 1.9.3     ✔ tidyr     1.3.1
## ✔ purrr     1.0.2     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

2. Make ggplots with mpg

mpg: data about cars that is already in Rcloud

using

ggplot(data = mpg)

we get an empty plot, because we haven’t told R what to do with the data

telling R what to do with the data

AFTER telling R that ggplot(data = mpg), we add onto it and use geom_point & aes() to tell R what to put on the plot

• geom_point(): creates a scatterplot with x and y axis that best shows a relationship between two continuous variables

• aes():decides what goes where on the scatterplot

• for this example we’ll be using the displ & hwy variables

picture <- ggplot(data = mpg) + 
  geom_point( # plots data points
    mapping = aes( # information about which data to analyse
      x = displ, #x axis information
      y = hwy #y axis information
      )
    )
print(picture)

• outcome looks like this

Appearance of the plot

1. Changing size of the points

Use size = cyl to make the size of the points vary depending on the variable:

picture <- ggplot(data = mpg) + 
  geom_point(
    mapping = aes(
      x = displ, 
      y=hwy,
      size = cyl # size is according to how many cylinders the car has
    )
  )
print(picture)

2. Changing the colour of the points

Use color = factor(cyl) to make the points of the colour change depending on the variable:

picture <- ggplot(data = mpg) + 
  geom_point(
    mapping = aes(
      x = displ, 
      y=hwy,
      color = factor(cyl)# to make the colours vary with level, not continuous
    )
  )
print(picture)

Use Color = “what colour I want” outside of aes() to make it a certain colour

• making it blue:

picture <- ggplot(data = mpg) + 
  geom_point(
    mapping = aes(
      x = displ, 
      y=hwy
      ),
    color = "blue"
  )
print(picture)

3. Adding a regression line/ rug

• Regression function: geom_smooth()
• Rug function: geom_rug() for margin distributions
• they can be added on top of all the code above:

picture <- ggplot(data = mpg) + 
  geom_point(
    mapping = aes(
      x = displ, 
      y=hwy,
      color = factor(cyl)
    )
  )+ #to add on another function
   geom_smooth(
    #for making the regression line onto the chart
    mapping = aes(
      x = displ,
      y = hwy
    )
  )+
  geom_rug(
    # geom_rug() is for margin distributions
    mapping = aes(
      x = displ,
      y = hwy,
      color = factor(cyl) # to make the margin colours according to cylinders
    )
  )
print(picture)

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

Cleaning the code

We want to avoid having redundancy in the code to make it as efficient as possible

Using the example above, we can change the factors x and y to universal factors by doing this:

picture <- ggplot(data = mpg,
       mapping = aes(
         x = displ,
         y = hwy # having them in the ggplot() bracket instead of the rest simplifies the information
       )
      )+
  geom_point(
    mapping = aes(
      color = factor(cyl)
    )
  )+
  geom_smooth()+
  geom_rug(
    mapping = aes(
      color = factor(cyl)
    )
  )
plot(picture) # for plotting, plot() and print() work the same way!

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

The output looks the same!

3. Making a boxplot

Boxplots are good for when you have one continuous and one discrete variable

load packages and csv

library(tidyverse)
forensic <- read.csv("data_forensic.csv")

Make the boxplot

We use geom_boxplot to turn data into boxplots

fbox <- ggplot(data = forensic,
               mapping = aes(
                x = band,
                y = est
               )
)+
  geom_boxplot(mapping = aes(
    fill = band #to make the boxes coloured according to band
  )
)+
  facet_wrap(vars(handwriting_expert)) + #groups the plots by handwriting expert/ novice
   theme_minimal() + # changes how the plot looks
  scale_x_discrete(
    name = NULL, # takes away the name (we already have it in the key)
    labels = NULL
  ) +
  scale_y_continuous(
    name = "Estimate" # gives the Y axis a clearer name
  )+
  ggtitle(
    label = "Handwriting feature probability for experts/ novices",
    subtitle = "Source: Martire et al."
  )+ # adds title/ subtitle to the whole plot
  scale_fill_viridis_d(
    alpha = .5, # transparency of the plot colours (to make median easier to find) 
    name = NULL # taking out the name of the keys
   )
print(fbox)            

## Warning: Removed 4 rows containing non-finite outside the scale range
## (`stat_boxplot()`).