Extension: Comparing City and Highway MPG

My extension adds tidyr::pivot_longer(). I use it to place city and highway MPG in the same column, then compare both values in one table and chart.

Load the data

Show code

# tidyverse includes dplyr, tidyr, and ggplot2.
# knitr is used to display the summary table.
library(tidyverse)
library(knitr)

# The mpg dataset is included with ggplot2, so no separate CSV is needed.
data("mpg", package = "ggplot2")

Change the table from wide to long

The original data stores city MPG in cty and highway MPG in hwy. I used pivot_longer() so both measurements use one road_type column and one mpg column.

Show code

# Keep the vehicle class and the two MPG columns.
# Rename cty and hwy so the labels are easier to understand.
mpg_long <- mpg |>
  select(class, city = cty, highway = hwy) |>
  # Stack city and highway into one road-type column.
  pivot_longer(
    cols = c(city, highway),
    names_to = "road_type",
    values_to = "mpg"
  ) |>
  mutate(
    class = recode(
      class,
      "2seater" = "Two Seater",
      "compact" = "Compact",
      "midsize" = "Mid-Size",
      "minivan" = "Minivan",
      "pickup" = "Pickup",
      "subcompact" = "Subcompact",
      "suv" = "SUV"
    ),
    road_type = str_to_title(road_type)
  )

mpg_long |>
  slice_head(n = 8) |>
  kable(
    col.names = c("Vehicle class", "Road type", "MPG"),
    caption = "City and highway MPG in a longer table"
  )

City and highway MPG in a longer table
Vehicle class	Road type	MPG
Compact	City	18
Compact	Highway	29
Compact	City	21
Compact	Highway	29
Compact	City	20
Compact	Highway	31
Compact	City	21
Compact	Highway	30

The row count changes from 234 to 468 because every vehicle now has one city row and one highway row. No vehicle was added or removed.

Calculate the averages

Show code

# Calculate one average for each vehicle class and road type.
mpg_summary <- mpg_long |>
  group_by(class, road_type) |>
  summarise(
    vehicles = n(),
    average_mpg = round(mean(mpg), 1),
    .groups = "drop"
  ) |>
  arrange(class, road_type)

mpg_summary |>
  kable(
    col.names = c("Vehicle class", "Road type", "Vehicles", "Average MPG"),
    caption = "Average city and highway MPG by vehicle class"
  )

Average city and highway MPG by vehicle class
Vehicle class	Road type	Vehicles	Average MPG
Compact	City	47	20.1
Compact	Highway	47	28.3
Mid-Size	City	41	18.8
Mid-Size	Highway	41	27.3
Minivan	City	11	15.8
Minivan	Highway	11	22.4
Pickup	City	33	13.0
Pickup	Highway	33	16.9
SUV	City	62	13.5
SUV	Highway	62	18.1
Subcompact	City	35	20.4
Subcompact	Highway	35	28.1
Two Seater	City	5	15.4
Two Seater	Highway	5	24.8

The table shows that every class has a higher highway average than city average. Compact and subcompact vehicles have the highest highway averages in this dataset.

Compare both road types in one chart

Show code

# The two road types use charcoal and orange, with printed values for accessibility.
# The value labels also show the difference without depending only on color.
ggplot(
  mpg_summary,
  aes(
    x = class,
    y = average_mpg,
    fill = road_type
  )
) +
  geom_col(
    position = position_dodge(width = 0.75),
    width = 0.68,
    color = "#333333"
  ) +
  geom_text(
    aes(label = average_mpg),
    position = position_dodge(width = 0.75),
    hjust = -0.15,
    size = 3.2
  ) +
  coord_flip() +
  scale_fill_manual(values = c("City" = "#666666", "Highway" = "#E69F00")) +
  scale_y_continuous(expand = expansion(mult = c(0, 0.15))) +
  labs(
    title = "Average city and highway MPG by vehicle class",
    x = NULL,
    y = "Average miles per gallon",
    fill = "Road type"
  ) +
  theme_minimal(base_size = 11) +
  theme(panel.grid.major.y = element_blank())

Grouped bar chart comparing average city and highway miles per gallon for seven vehicle classes. Highway MPG is higher for every class. — Figure 1

Summary

My extension adds a new Tidyverse step to the original work. pivot_longer() makes it possible to group city and highway values under one variable. This makes the comparison easier to summarize and graph without writing separate code for each road type.

Sources

AI use

I used ChatGPT and VS Code to help draft, troubleshoot, and organize parts of the code and documentation.