ggplot2 grammar of graphics

ggplot2 is R's signature plotting library, built on the Grammar of Graphics. Once you internalise the grammar — data + aesthetics + geom + scales + themes — you can build virtually any chart with a small consistent vocabulary.

The grammar

library(ggplot2)

ggplot(bankrates, aes(x = month, y = lending_rate)) +
    geom_line() +
    labs(title = "Kenyan lending rates", y = "Rate (%)")

Every ggplot has at minimum: data (the data frame), an aes mapping (which columns control which visual properties), and one or more geoms (the geometric shapes — point, line, bar, etc.). You add layers with +.

Common geoms

• geom_point — scatter plot
• geom_line — line plot
• geom_bar — bar plot (counts)
• geom_col — bar plot (values)
• geom_histogram — histogram
• geom_boxplot — boxplot
• geom_smooth — fitted regression curve with CI

Aesthetics in detail

ggplot(bankrates, aes(x = month, y = lending_rate, colour = bank)) +
    geom_line()
# colour = bank means draw a separate line per bank, with auto-legend

Faceting — small multiples

ggplot(bankrates_long, aes(x = month, y = rate)) +
    geom_line() +
    facet_wrap(~ bank, scales = "free_y")
# One panel per bank, each with its own y-axis scale

Scales and themes

ggplot(bankrates, aes(x = month, y = lending_rate)) +
    geom_line() +
    scale_y_continuous(labels = scales::percent) +
    theme_minimal() +
    labs(title = "Lending rates", x = NULL, y = "Rate")