From data to plot in one pipeline in R

The tidyverse habit: reshape, then plot, without an intermediate variable.

library(dplyr)
library(tidyr)
library(ggplot2)

wide <- tibble(
  lang = c("r", "python", "sql"),
  week_1 = c(88, 99, 70),
  week_2 = c(92, 104, 74),
  week_3 = c(96, 101, 79)
)

plot <- wide |>
  pivot_longer(starts_with("week"), names_to = "week", values_to = "tpm",
               names_prefix = "week_", names_transform = list(
                 week = as.integer)) |>
  mutate(improvement = tpm - first(tpm), .by = lang) |>
  ggplot(aes(week, tpm, colour = lang)) +
  geom_line(linewidth = 1) +
  geom_point(size = 2) +
  scale_x_continuous(breaks = 1:3) +
  labs(title = "Three weeks", y = "tpm", x = "week") +
  theme_minimal()

print(nrow(plot$data))
ggsave(tempfile(fileext = ".png"), plot, width = 5, height = 3, dpi = 100)

How it works

1. The pipe carries the data into ggplot.
2. + continues the plot; |> continues the data.
3. One expression, so nothing half-built escapes.

Keywords and builtins used here

• aes
• c
• first
• geom_line
• geom_point
• ggplot
• ggsave
• labs
• library
• list
• mutate
• nrow
• pivot_longer
• print
• scale_x_continuous
• starts_with
• tempfile
• theme_minimal
• tibble

The run, in numbers

Lines

25

Characters to type

677

Tokens

207

Three-star pace

105 tpm

At the three-star pace of 105 tokens a minute, this run takes about 118 seconds.

Type this snippet

Step 3 of 3 in Finishing a plot, step 14 of 15 in ggplot2.

← Previous Next →