• The read_* functions are extremely clever, helping you read data and always returning a tibble.

• A tibble is a “nicer” and more honest data.frame

• Will this turn out as you imagined?

data.frame(x=list("="="=", "B"="b"), # length 2
           y=c(1,2,3)) # length 3
tibble(x=list("="="=", "B"="b"),
       y=c(1,2,3))

data.frame(x=list("A"="a", "B"="b"),
           y=list("C"="c", "D"="d"))
tibble(x=list("A"="a", "B"="b"),
       y=list("C"="c", "D"="d"))

• Faster than base-R read.*, used in the same way.
• Don’t make strings into factors.
• Nothing strange about your column names.
• Does not depend on environment variables in your system, good for reproducibility among machines/systems.
• Useful arguments col_types (colClasses from read.*) and guess_max
• We will complain if you use read.csv instead of read_csv

• Package readxl, excel files are special
• readxl::read_excel function to read .xls/ .xlsx files.
• Unique to read_excel the sheet argument.
• Will be used in Homework 2.

mtcars <- tibble::rownames_to_column(mtcars, var = "model")
kable(head(arrange(mtcars, mpg), n = 4))

kable(head(arrange(mtcars, mpg, disp), n = 4))

#only those with manual transmission
kable(head(filter(mtcars, am == 1), n=4))

kable(head(filter(mtcars, mpg < 30), n=4))

kable(head(mutate(mtcars, lpm = 235 / mpg), n=4))

kable(head(select(mtcars, model, mpg), n=4))

What function does the argument x=3 take?

df <- f1(f2(f3(f4(mtcars, x=2), x), x=3))

This is not easy to read! How can we solve it?

Determine \(h\circ g \circ f(a) = h(g(f(a)))\)

Three different ways to calculate this in R:

b <- f(a)
c <- g(b)
h(c)

h(g(f(a)))

a %>%
    f() %>%
    g() %>%
    h()

mtcars <- mutate(mtcars, lpm = 235 / mpg)
mtcars <- filter(mtcars, am == 1)
ggplot(mtcars, aes(x = hp, y = lpm)) + geom_point()

ggplot(
    filter(
        mutate(mtcars, lpm = 235 / mpg)
        , am ==1),
    aes(x = hp, y = lpm)) + geom_point()

mtcars %>%
    mutate(lpm = 235 / mpg) %>%
    filter(am == 1) %>%
    ggplot(aes(x = hp, y = lpm)) + geom_point()
    

A statistical plot has components

• data
• geom: type of geometric objects (points, lines, …)
• cord: coordinate system
• mapping: binds data to the dimensions/“aesthetics” of the coordinate system (position, color, shape, size, …)

A scatterplot

• data: mpg and hp for a number of cars
• geom: points
• coord: Cartesian
• mapping: binds hp to position on the x axis and mpg to the y axis

ggplot(data = mtcars, mapping = aes(x = hp, y = mpg)) + geom_point()

ggplot(mtcars,
       aes(x = hp, y = mpg, size = wt, color = cyl)) +
    geom_point()

What is cyl?

ggplot(mtcars,
       aes(x = hp, y = mpg, size = wt, color = as.factor(cyl))) +
    geom_point()

ggplot(mtcars,
       aes(x = hp, y = mpg, size = wt)) +
    geom_point(color = cyl)
# ERROR

ggplot(mtcars,
       aes(x = hp, y = mpg, size = wt)) +
    geom_point(color = "red")
# OK

aes looks in mtcars but not geom!

Be careful about reusing names in columns, variables, etc.

cyl <- "blue"
ggplot(mtcars,
       aes(x = hp, y = mpg, size = wt)) +
    geom_point(color = cyl)

Error?