Draw and customize a bar chart with base R barplot()
Draw a bar plot in base R with barplot(). Chart counts from a table or values from a named vector, rename and colour the bars, go horizontal, add value labels, and build stacked and grouped bar charts from a matrix with a legend.
barplot() — no package needed. Pass a named vector of values, or a table() of counts.names.arg, add a heading and axis titles with main, xlab, ylab.col and the outline with border; pass a vector to colour each bar differently.horiz = TRUE; add value labels by capturing the bar midpoints bp <- barplot(...) and calling text(bp, …).barplot() a matrix to get stacked bars, or add beside = TRUE for grouped bars — then label them with legend().A bar plot turns a set of category totals into a row of bars whose heights you can compare at a glance — counts per group, a value per category, a survey tally. In base R you draw one with barplot(), and it takes two kinds of input: a named vector of values (one bar per name) or a table() of counts (one bar per level). Neither needs a package.
This lesson covers both inputs, then the styling you actually reach for — renaming bars, titles, fill and border colours, horizontal layout, value labels on top of the bars — and finishes with the two multi-group forms: stacked and grouped bars from a matrix, tied together with a legend().
We use two built-in datasets: mtcars, where table(mtcars$cyl) counts cars by cylinder number, and VADeaths, a matrix of 1940 Virginia death rates (per 1000) cross-classified by age band and population group.
# Counts: how many cars have 4, 6, or 8 cylinders?
table(mtcars$cyl)
4 6 8
11 7 14 # Values: a matrix of death rates (age band x population group)
VADeaths Rural Male Rural Female Urban Male Urban Female
50-54 11.7 8.7 15.4 8.4
55-59 18.1 11.7 24.3 13.6
60-64 26.9 20.3 37.0 19.3
65-69 41.0 30.9 54.6 35.1
70-74 66.0 54.3 71.1 50.0Prefer the grammar of graphics? ggplot2’s geom_bar() and geom_col() build the legend and theming for you — see the ggplot2 series.
The most common bar plot is a count per category. table() tallies a factor, and barplot() draws one bar per level. Here is the number of cars at each cylinder count, filled in brand azure:
counts <- table(mtcars$cyl)
barplot(counts,
col = "#3a86d4", border = "white",
main = "Cars by cylinder count",
xlab = "Cylinders", ylab = "Number of cars")
table(mtcars$cyl) supplies both the heights and the bar labels (4, 6, 8). col fills the bars and border = "white" gives them a clean separating edge; main, xlab and ylab set the title and axis labels — barplot() draws no axis titles by default.
When you already have one value per category — not raw data to count — pass a named vector. Each element becomes a bar and its name becomes the label. We pull one column out of VADeaths, which gives a named vector of death rates by age band:
deaths <- VADeaths[, "Rural Male"]
deaths50-54 55-59 60-64 65-69 70-74
11.7 18.1 26.9 41.0 66.0 barplot(deaths,
col = "#3a86d4", border = "white",
main = "Rural male death rates",
xlab = "Age band", ylab = "Deaths per 1000")
deaths is a named numeric vector, so barplot() uses the values as heights and the names (50-54, 55-59, …) as labels. The rates climb steadily with age, exactly as you’d expect.
Override the labels with names.arg — a vector with one label per bar, in order. This is handy when the names are long, coded, or you want cleaner text than the data carries:
deaths <- VADeaths[, "Rural Male"]
barplot(deaths,
names.arg = c("A", "B", "C", "D", "E"),
col = "#3a86d4", border = "white",
main = "Rural male death rates",
xlab = "Age group", ylab = "Deaths per 1000")
names.arg must list one label per bar, in the same order as the values. Everything else — heights, colour, titles — is unchanged.
Set horiz = TRUE to lay the bars on their side. Horizontal bars read well when the category labels are long, or when you simply want the value axis along the bottom:
deaths <- VADeaths[, "Rural Male"]
barplot(deaths,
horiz = TRUE, las = 1,
col = "#3a86d4", border = "white",
main = "Rural male death rates",
xlab = "Deaths per 1000")
With horiz = TRUE the value axis moves to the bottom and the categories stack up the side. las = 1 turns the axis labels upright so the age bands stay readable.
col fills the bars; border colours their outlines. Give either a single colour for every bar, or a vector to colour each bar differently. A vector is useful when the colour itself carries meaning — here a colourblind-safe viridis set from base R’s hcl.colors():
op <- par(mfrow = c(1, 2))
deaths <- VADeaths[, "Rural Male"]
bar_cols <- hcl.colors(5, "viridis") # colourblind-safe, base R
# Left: a single fill colour
barplot(deaths, col = "#3a86d4", border = "white",
main = "Single fill")
# Right: one colour per bar
barplot(deaths, col = bar_cols, border = "white",
main = "Colour per bar")
par(op)Use one colour when the bars are just categories you compare by height; use a vector — a colourblind-safe palette like hcl.colors(n, "viridis") — when the colour is meant to distinguish the categories.
barplot() returns the x-positions of the bar centres — capture them in a variable and you can place the exact value on top of each bar with text(). This is the base-R idiom for value labels:
counts <- table(mtcars$cyl)
bp <- barplot(counts,
col = "#3a86d4", border = "white",
ylim = c(0, max(counts) * 1.15), # headroom for the labels
main = "Cars by cylinder count",
xlab = "Cylinders", ylab = "Number of cars")
text(x = bp, y = counts, labels = counts,
pos = 3, font = 2, col = "#1d4e7a")
bp holds the bar midpoints, so text(x = bp, y = counts, …) writes each count at the top of its bar; pos = 3 nudges the label just above the bar and ylim adds headroom so it isn’t clipped. The labels sit on the white background, so they stay easy to read.
Hand barplot() a matrix and it stacks the rows within each column. VADeaths is already a matrix — each column is a population group, each row an age band — so one call gives four stacked bars. Pass legend = rownames(...) to label the stacked segments:
bar_cols <- hcl.colors(5, "viridis")
barplot(VADeaths,
col = bar_cols, border = "white",
legend = rownames(VADeaths),
main = "Death rates by population group",
ylab = "Deaths per 1000")
Each bar is one column of VADeaths, and the five age-band rows stack from bottom to top. The colour vector maps to the rows, and legend = rownames(VADeaths) keys the age bands to their colours. Stacking shows the total per group and the composition inside it at the same time.
Add beside = TRUE and the same matrix is drawn as grouped bars instead — one cluster per column, one bar per row. This is the form to reach for when you want to compare the individual age bands across groups, rather than read totals:
bar_cols <- hcl.colors(5, "viridis")
barplot(VADeaths,
beside = TRUE,
col = bar_cols, border = "white",
legend = rownames(VADeaths),
main = "Death rates by age band",
ylab = "Deaths per 1000")
beside = TRUE places the five age-band bars side by side within each group. Now you can see at a glance that the rate rises with age in every group and that the urban and male groups run higher — a comparison the stacked version hides.
The built-in legend = argument sometimes overlaps the tallest bars. For full control, drop it and add a separate legend() call, which lets you choose the position and styling:
bar_cols <- hcl.colors(5, "viridis")
barplot(VADeaths,
beside = TRUE,
col = bar_cols, border = "white",
main = "Death rates by age band",
ylab = "Deaths per 1000")
legend("topleft",
legend = rownames(VADeaths),
fill = bar_cols,
box.lty = 0, cex = 0.8)
legend("topleft", …) puts the key where you want it; fill = bar_cols draws the matching colour swatches. box.lty = 0 removes the box around the legend and cex = 0.8 shrinks the text so it doesn’t crowd the bars. See Add a legend for the full legend() toolkit.
Both charts draw the same matrix; they answer different questions.
| Form | Argument | Best for |
|---|---|---|
| Stacked | default (a matrix) | The total per group, and each segment’s share of it. |
| Grouped | beside = TRUE |
Comparing the individual categories across groups. |
Reach for stacked when the whole matters (a total split into parts); reach for grouped when you need to compare like-for-like bars side by side.
Swap the dataset, switch beside = TRUE on or off, or change the colours and press Run — it executes in your browser via webR, no install needed.
Ask Prova “how do I put the value on top of each bar in a base R barplot?” — it answers with base-R code you can run, so the explanation is reproducible. The runtime is the judge. Ask Prova →
Only one bar appears when I pass a whole data frame or two columns. barplot() wants a vector or a matrix, not a data frame. For counts, tabulate first — barplot(table(mtcars$cyl)). For a values chart, pass a single named vector — barplot(VADeaths[, "Rural Male"]). For multiple series, pass a matrix (as.matrix(df)), which triggers the stacked/grouped behaviour.
My value labels get cut off at the top of the plot. The default y-axis stops at the tallest bar, so a label sitting above it is clipped. Add headroom with ylim = c(0, max(counts) * 1.15) before writing the labels with text(..., pos = 3).
Call barplot() on a named numeric vector (one bar per name) or a table() of counts (one bar per level) — for example barplot(table(mtcars$cyl)). No package is required; it ships with base R. Add col for the fill and main/xlab/ylab for the labels.
Tabulate the variable first, then plot the table: barplot(table(mtcars$cyl)). table() counts how many observations fall in each category and supplies both the bar heights and the labels, so barplot() draws one bar per level automatically.
Give barplot() a matrix. By default the rows are stacked within each column; add beside = TRUE for grouped (side-by-side) bars. Pass legend = rownames(m) and a col vector to colour and label the series.
Add horiz = TRUE to the barplot() call. The value axis moves to the bottom and the categories stack up the side; add las = 1 to keep the category labels upright and readable.
Capture the bar positions that barplot() returns — bp <- barplot(x) — then call text(x = bp, y = x, labels = x, pos = 3). Give the plot headroom with ylim = c(0, max(x) * 1.15) so the labels above the tallest bar aren’t clipped.
Starting from the VADeaths matrix, draw a grouped bar plot (bars side by side, not stacked) using the five colourblind-safe colours below, and add a legend in the top-left corner labelling the age bands.
Pass beside = TRUE and col = bar_cols to barplot(). Then call legend("topleft", legend = rownames(VADeaths), fill = bar_cols).
bar_cols <- hcl.colors(5, "viridis")
barplot(VADeaths, beside = TRUE,
col = bar_cols, border = "white",
main = "Death rates by age band",
ylab = "Deaths per 1000")
legend("topleft",
legend = rownames(VADeaths),
fill = bar_cols, box.lty = 0, cex = 0.8)beside = TRUE clusters the bars; legend(fill = bar_cols) draws swatches that match the bar colours.
Quick check. You call barplot(VADeaths) with no other arguments. Are the bars stacked or grouped?
Stacked. For a matrix, barplot() stacks the rows within each column by default. You get grouped (side-by-side) bars only when you add beside = TRUE.
barplot() is the one function you need for bar charts in base R. Pass a named vector for a values chart, or a table() for counts. From there, names.arg relabels the bars, col/border set the colours, horiz = TRUE flips the layout, and capturing the returned midpoints lets text() add value labels. Give it a matrix for stacked bars, add beside = TRUE for grouped bars, and tie the series together with legend(). That covers the everyday bar plot end to end.
@online{2026,
author = {},
title = {Bar {Plot} in {R} {Base} {Graphics}},
date = {2026-07-08},
url = {https://www.datanovia.com/learn/data-visualization/r-base/barplot},
langid = {en}
}