Agenda

Data visualisation in R for Data Science, Section 3.5.1.

Do first problem set
Read in data file
Plotting facets
Plotting curves
Combining plot types

Do First problem set

Before each new session, we'll do a quick recap, called a Do First. These will only use functions we've previously covered, so if you're unsure or can't remember, just check the code from the previous session.

Recreate the below plot using the smaller NYC Airbnb dataset (nyc from Lesson 1). There are four aesthetics to change and the plot uses theme_solarized.
Hint: Use the help ? function if something isn't clear.

# You didn't think we'd make it this easy, did you?

Some useful shortkeys for making `R` life easier

TAB = autofill rest of function/global variable
CTRL + ENTER = run code
ALT + minus sign = insert assign operator <-
CTRL + SHIFT + M = insert pipe %>%
Run ALT + SHIFT + K for all available shortkeys

Read in data

my_file <- "your_csv_file.csv"
my_data <- read_csv(my_file)  # read in the csv data file  
glimpse(my_data)

Grouping data

One way to group your data is by colour

my_data <- mpg
my_theme <- theme_classic()
ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy, colour = class)) + my_theme

Plotting facets

facet_wrap and facet_grid
Facets add a third variable to a plot
The facet function takes a formula as an argument, which is just a data structure, denoted by a tilde ~

facet_wrap
When you have one variable to plot as a facet

ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + facet_wrap(~class, nrow = 2) + 
    my_theme

facet_grid
When you know the two variables you want to plot
The formula structure for facet_grid is Y variable ~ X variable, e.g. drv ~ cyl

ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + facet_grid(drv ~ cyl) + my_theme

You can also replace the X or Y argument in facet_grid with a period (".") to plot only one variable.

# Y var
ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + facet_grid(. ~ cyl) + my_theme

# X var
ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + facet_grid(drv ~ .) + my_theme

Exercise 3.5.1

Try the exercises from 3.5.1.

`geom_smooth`

Plotting points (geom_point) or lines (geom_smooth)

# left
ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + my_theme

# right
ggplot(data = my_data) + geom_smooth(mapping = aes(x = displ, y = hwy)) + my_theme

Linetype

Grouping by linetype

ggplot(data = my_data) + geom_smooth(mapping = aes(x = displ, y = hwy, linetype = drv)) + my_theme

Group vs. colour

Using group separates the data into objects ...

ggplot(data = my_data) + geom_smooth(mapping = aes(x = displ, y = hwy)) + my_theme

ggplot(data = my_data) + geom_smooth(mapping = aes(x = displ, y = hwy, group = drv)) + my_theme

... but colour will distinguish the differences among these objects

ggplot(data = my_data) + geom_smooth(mapping = aes(x = displ, y = hwy, colour = drv), show.legend = FALSE) + 
    my_theme

Geometric objects

Adding `geoms`

Possibly the most useful part of plotting data is layering plot types

ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + geom_smooth(mapping = aes(x = displ, 
    y = hwy)) + my_theme

# condensing code
ggplot(data = my_data, mapping = aes(x = displ, y = hwy)) + geom_point() + geom_smooth() + my_theme

# adding aes
ggplot(data = my_data, mapping = aes(x = displ, y = hwy)) + geom_point(colour = "steel blue") + geom_smooth(colour = "#C6BDEA", 
    fill = "#C6BDEA") + my_theme

But why does this throw an error?

# adding aes
ggplot(data = my_data) + geom_point(mapping = aes(x = displ, y = hwy)) + geom_smooth() + my_theme

Error: stat_smooth requires the following missing aesthetics: x and y

Specifying layers

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

Applying different datasets to one plot (overriding data)

require(dplyr)
names(my_data)

 [1] "manufacturer" "model"        "displ"        "year"         "cyl"          "trans"       
 [7] "drv"          "cty"          "hwy"          "fl"           "class"

# subset data with filter
my_data_subcompact <-  filter(filter(my_data, class == "subcompact"))

ggplot(data = my_data, mapping = aes(x = displ, y = hwy)) + 
  geom_point(mapping = aes(color = class)) + # original data
  geom_smooth(data = my_data_subcompact, se = FALSE) + # filtered data 
  my_theme

Exercise 3.6.1

Try the exercises from 3.6.1.

# 1
ggplot(my_data)  # ...

# 2

Applying the Airbnb data

Use the new examples on the Airbnb dataset.

Lesson 2: Reading data and plotting facets and curves