This presentation was made in RStudio using RMarkdown!

• Can set output: ioslides_presentation in the header

• Anybody not on Slack/DataCamp yet?
• Anybody not filled out the background survey?
• Anybody not signed up to meet with me?
• Anybody not have a working RStudio server account?

• R commands are like sentences:

  • Functions are like verbs
    
  • Some arguments are like nouns (with particular roles)
    
  • (Other arguments are like prepositional phrases)

draw(picture, recipient = "me", material = "paint")

• takes the thing on the left, and (by default) puts it (or its output, if it is a function call) into the first argument slot for the function on the right

• We'll do a lot more with this later

## Two equivalent expressions
arbuthnot %>% mutate(total = boys + girls)
mutate(arbuthnot, total = boys + girls)

• Takes a data frame as the main (first) argument
• Takes a "formula" defining a new variable as the second argument
• Output is a new data frame with a new column added
• If we want to keep the new data frame, need to assign it to a named "container"

## Creates the column, but it immediately disappears
mutate(arbuthnot, total = boys + girls)
## Creates the column, and creates a brand new data frame
## that has both new and old variables (now we have two
## data frames, one with and one without the new variable)
new.arbuthnot <- mutate(arbuthnot, total = boys + girls)
## Creates the column, and overwrites the original data
## with a data frame that has new and old columns
arbuthnot <- mutate(arbuthnot, total = boys + girls)

• Visual cues
  • position, size, color, etc.
• Coordinate system
  • how are data points organized?
• Scale
  • relationship between variable and distance in space
• Context
  • what in the world is the data about?
• Faceting
  • What are the sub-parts (facets) of the graph?

Source: Nathan Yau, Data Points

Source: Nathan Yau, Data Points

Source: Nathan Yau, Data Points

Source: Nathan Yau, Data Points

Source: Nathan Yau, Data Points

Source: New York Times

Cleveland and McGill (1985): people better at judging:

• position than size
• length than angle
• 1D differences than 2D differences
• 2D differences than 3D differences
• size than color

• Prefer length scale to color scale
• Never use pie charts
• Always Bring Context
• Above all else, show the data (Edward Tufte)

• The Grammar of Graphics is a book by Leland Wilkinson (1999, 2005) that set out to define the "parts of speech" and "grammar rules" of data visualization
• An "ontology of graphs"
• Implemented in R in ggplot2 (Hadley Wickham, 2010)
  • ggplot (or ggplot1) was sort of a beta version; not really used today

The basic operation is combination of these elements via the + operator. A fully formed combination returns a plot as an R object (can be assigned to things, operated on later, etc.)

Plots must at a minimum have data (data=), a mapping (aes()), and at least one geometry element (geom_*())

library(tidyverse)
ggplot(data = mtcars, aes(x = disp, y = mpg)) +
  geom_point()

ggplot(data = mtcars, aes(x = disp, y = mpg, color = factor(cyl))) +
  geom_point() + 
  geom_line() + 
  facet_wrap(~am) +
  scale_color_brewer(palette = "Set1")