19 Data Reshaping for ggplot2

In this chapter, we will learn the one idea that unlocks faceting, grouped aesthetics and multi-series line graphs: long versus wide data.

Why this matters

ggplot2 maps one column to one aesthetic. Wide tables (one column per series) must be reshaped to long tables (one key column plus one value column) before color, fill or facets can separate the series.

Common pitfall

gather() appears in older answers but is retired. Always use tidyr::pivot_longer() in new code.

19.1 Wide Data

Wide data keeps each series in its own column. Here quarterly revenue for two regions sits in north and south:

wide <- data.frame(
  quarter = c("Q1", "Q2", "Q3", "Q4"),
  north = c(120, 150, 135, 180),
  south = c(100, 130, 160, 175)
)
wide

  quarter north south
1      Q1   120   100
2      Q2   150   130
3      Q3   135   160
4      Q4   180   175

Plotting this directly forces one layer per series — repetitive and unscalable.

19.2 Long Data

pivot_longer() collapses the series columns into a region key column and a revenue value column:

long <- tidyr::pivot_longer(wide, cols = c(north, south),
  names_to = "region", values_to = "revenue")
long

# A tibble: 8 × 3
  quarter region revenue
  <chr>   <chr>    <dbl>
1 Q1      north      120
2 Q1      south      100
3 Q2      north      150
4 Q2      south      130
5 Q3      north      135
6 Q3      south      160
7 Q4      north      180
8 Q4      south      175

19.3 Plotting Long Data

One mapping now draws both series, distinguished by color:

ggplot(long) +
  geom_line(aes(quarter, revenue, color = region, group = region), linewidth = 1) +
  geom_point(aes(quarter, revenue, color = region), size = 2)

The same long shape powers facets and dodged bars with no extra reshaping.

19.4 Modern Pipe

Chain reshaping into plotting with the native pipe |>:

wide |>
  tidyr::pivot_longer(cols = c(north, south), names_to = "region", values_to = "revenue") |>
  ggplot() +
  geom_bar(aes(region, revenue, fill = quarter), stat = "identity",
    position = position_dodge(width = 0.8))

Where to go next

tidyr pivot_longer() reference for multi-column specs
Line Graphs chapter for multi-series line recipes
Faceting chapter for small multiples of long data