How to Draw a Box and Whisker Diagram: Complete Step-by-Step Guide (2026)

Drawing a Box and Whisker Diagram: A Full Tutorial for Researchers and Students

Picture a column of numbers with no obvious pattern. How do you convey, in a single picture, where the bulk of those numbers sit, how widely they scatter, and whether a few of them stray far from the pack? A box and whisker diagram answers all three questions at once, using a footprint smaller than most paragraphs.

The format traces back to John Tukey, who introduced it in 1969. Today it travels under several interchangeable names: box plot, box and whisker plot, box and whisker diagram. Whatever you call it, the graphic distills a distribution down to five summary numbers. That economy explains why it turns up constantly in lab reports, journal figures, and lecture slides across ecology, neuroscience, mechanical engineering, economics, and beyond, wherever someone needs to line up several groups and ask how they differ.

The sections below walk through every route to the same chart, beginning with paper and a pencil and ending with reproducible scripts in Excel, Python, and R.

AI Chart Generator

Generate professional box plots and statistical charts from your data using AI. Export in high resolution for research papers and presentations.

Create box plots free →

What a Box and Whisker Diagram Actually Shows

At its core, the chart encodes a distribution through five reference points: the lowest value, the first quartile (Q1), the median (Q2), the third quartile (Q3), and the highest value. Taken as a set, these numbers reveal three things simultaneously, where the data tends to center, how loosely or tightly it spreads, and whether stray observations sit far from everything else.

Compare that against the usual alternatives. A bar chart collapses everything into an average and hides the variation behind it. A histogram brings the shape back but forces you to pick a bin width that can change the story. The box plot threads the needle: it gives a tidy, bin-free portrait of spread that stays legible even when you stack a dozen groups along one axis.

Why Box Plots Appear Throughout Academic Research

A few concrete payoffs keep box plots in heavy rotation among researchers:

We expand on how chart selection underpins clear scientific reporting in our research data visualization guide, where box plots repeatedly stand out as the natural pick for side-by-side distribution comparisons.

The Anatomy of a Box and Whisker Plot

Once you can name each part of the chart and say what it stands for, both drawing one and reading one become almost mechanical.

The Five-Number Summary

These are the five quantities that every box plot is assembled from:

1. Minimum: The lowest reading that remains once outliers are set aside
2. First Quartile (Q1): The cut point at the 25th percentile, with a quarter of the observations lying beneath it
3. Median (Q2): The middle of the data at the 50th percentile, dividing the set into two equal halves
4. Third Quartile (Q3): The cut point at the 75th percentile, with three quarters of the observations lying beneath it
5. Maximum: The highest reading that remains once outliers are set aside

Visual Components Explained

A scatter plot lays bare the individual points that a box plot later folds into five summary numbers. Looking at the two side by side gives you the fullest read on your data.

How to Draw a Box and Whisker Diagram by Hand

Nothing cements the concept like running the arithmetic yourself. We will track a small operations dataset all the way through: the delivery times, in minutes, recorded for twelve orders from a campus food stall.

Dataset: 14, 17, 19, 21, 22, 24, 26, 27, 29, 31, 34, 58

Notice that this set has an even number of readings and one suspiciously slow delivery, which makes it a good stress test for the procedure.

Step 1: Sort the Data

Put every reading in ascending order first. Ours already arrives sorted, which keeps the bookkeeping simple:

14, 17, 19, 21, 22, 24, 26, 27, 29, 31, 34, 58

Step 2: Find the Median (Q2)

With twelve readings, no single value sits dead center. The median lands between the 6th and 7th entries, so we average them:

Q2 (Median) = (24 + 26) / 2 = 25

Step 3: Find Q1 (First Quartile)

Q1 is the median of the lower half, the six readings that fall below the overall median:

14, 17, 19, 21, 22, 24

Again the half has an even count, so Q1 is the average of its two middle entries:

Q1 = (19 + 21) / 2 = 20

Step 4: Find Q3 (Third Quartile)

Q3 is the median of the upper half, the six readings above the overall median:

26, 27, 29, 31, 34, 58

Q3 = (29 + 31) / 2 = 30

Step 5: Compute the IQR and Fence Values

IQR = Q3 - Q1 = 30 - 20 = 10

• Lower fence: Q1 - 1.5 x IQR = 20 - 15 = 5
• Upper fence: Q3 + 1.5 x IQR = 30 + 15 = 45

Here the fences earn their keep. The 58-minute delivery sits above the upper fence of 45, so it qualifies as an outlier and gets plotted on its own.

Step 6: Set the Whisker Endpoints

• Lower whisker endpoint: the smallest reading still at or above the lower fence = 14
• Upper whisker endpoint: the largest reading still at or below the upper fence = 34

Step 7: Draw the Chart

1. Lay down a number line that spans the entire range of readings
2. Sketch a rectangle stretching from Q1 (20) to Q3 (30)
3. Drop a vertical line inside the box at the median (25)
4. Run a whisker from the left edge of the box down to 14
5. Run a whisker from the right edge of the box out to 34
6. Mark the lone 58-minute reading as a separate point past the right whisker

Pulling it together, the five-number summary plus the flagged outlier reads:

Creating Box Plots in Excel

From Excel 2016 onward, box and whisker is a first-class chart type, so you can hand the spreadsheet your raw numbers and let it do the quartile math.

Using the Built-In Chart Option

1. Drop your values into one column, with a label in the top cell
2. Highlight that column, header included
3. Walk through Insert > Charts > Statistical > Box and Whisker
4. Excel works out the quartiles on its own and draws the finished chart

Formatting Options

Right-click the box itself and a panel of toggles appears:

• Show inner points: Plot every reading that lands inside the whiskers
• Show outlier points: Surface readings that sit past the whisker tips
• Show mean markers: Drop a marker at the arithmetic mean
• Show mean line: Trace a line connecting the means across groups
• Inclusive or Exclusive quartile calculation: Pick the formula Excel uses for Q1 and Q3

Getting Publication-Ready Output

Creating Box Plots in Python

When you want precise command over every line, fill, and marker, Matplotlib and Seaborn are hard to beat.

Basic Box Plot with Matplotlib

import matplotlib.pyplot as plt

data = [14, 17, 19, 21, 22, 24, 26, 27, 29, 31, 34, 58]

fig, ax = plt.subplots(figsize=(8, 5))
ax.boxplot(data, vert=False, patch_artist=True,
           boxprops=dict(facecolor='#4C72B0', alpha=0.7),
           medianprops=dict(color='white', linewidth=2))

ax.set_xlabel('Delivery Time (minutes)')
ax.set_title('Distribution of Order Delivery Times')
plt.tight_layout()
plt.savefig('boxplot.png', dpi=300)
plt.show()

Comparing Multiple Groups with Seaborn

import seaborn as sns
import pandas as pd

# Sample data for three treatment groups
data = pd.DataFrame({
    'Group': ['Control']*20 + ['Treatment A']*20 + ['Treatment B']*20,
    'Value': [45, 48, 42, 50, 47, 44, 49, 51, 46, 43,
              53, 48, 55, 50, 52, 47, 54, 51, 49, 56,
              60, 62, 58, 65, 61, 59, 63, 57, 64, 66,
              55, 52, 58, 53, 57, 54, 56, 59, 51, 60]
})

fig, ax = plt.subplots(figsize=(8, 6))
sns.boxplot(x='Group', y='Value', data=data, palette='Set2', ax=ax)
ax.set_ylabel('Measurement Value')
ax.set_title('Treatment Comparison')
plt.tight_layout()
plt.savefig('comparison_boxplot.png', dpi=300)
plt.show()

Overlaying Individual Data Points

A growing number of journals ask you to show the underlying observations on top of the summary box:

fig, ax = plt.subplots(figsize=(8, 6))
sns.boxplot(x='Group', y='Value', data=data, palette='Set2', ax=ax)
sns.stripplot(x='Group', y='Value', data=data,
              color='black', alpha=0.4, size=4, ax=ax)
plt.tight_layout()
plt.savefig('boxplot_with_points.png', dpi=300)
plt.show()

Creating Box Plots in R

Across many quantitative fields, R paired with ggplot2 is the default workshop for statistical figures.

Basic Box Plot with ggplot2

library(ggplot2)

deliveries <- data.frame(value = c(14, 17, 19, 21, 22, 24,
                                   26, 27, 29, 31, 34, 58))

ggplot(deliveries, aes(y = value)) +
  geom_boxplot(fill = "#4C72B0", alpha = 0.7, width = 0.4) +
  labs(title = "Distribution of Order Delivery Times",
       y = "Delivery Time (minutes)") +
  theme_minimal()

Group Comparisons in R

# Sample grouped data
treatment_data <- data.frame(
  Group = rep(c("Control", "Treatment A", "Treatment B"), each = 20),
  Value = c(rnorm(20, mean = 47, sd = 3),
            rnorm(20, mean = 52, sd = 4),
            rnorm(20, mean = 60, sd = 3))
)

ggplot(treatment_data, aes(x = Group, y = Value, fill = Group)) +
  geom_boxplot(alpha = 0.7) +
  geom_jitter(width = 0.15, alpha = 0.4, size = 2) +
  scale_fill_brewer(palette = "Set2") +
  labs(title = "Treatment Group Comparison",
       y = "Measurement Value") +
  theme_minimal() +
  theme(legend.position = "none")

Notched Box Plots for Visual Significance Testing

The notch carved around the median traces a rough 95% confidence interval. If the notches of two boxes fail to overlap, you have a quick visual hint that their medians genuinely differ:

ggplot(treatment_data, aes(x = Group, y = Value, fill = Group)) +
  geom_boxplot(notch = TRUE, alpha = 0.7) +
  scale_fill_brewer(palette = "Set2") +
  theme_minimal()

When Box Plots Are the Right Choice

For all their strengths, box plots are not a default for every dataset. The checklist below helps you decide when they earn their place.

Situations Where Box Plots Work Well

• You are contrasting two or more distributions on the same scale
• You need outliers in continuous measurements to jump out automatically
• You want spread captured compactly, median and quartiles and range together
• Several summary values matter at once and space is tight
• You are weighing median differences between experimental conditions

When Another Chart Type Fits Better

Our guide to scientific diagrams makes the same point: a figure works best when its type fits both the shape of the data and the people reading it.

A bar chart hands you the average and stops there. A box plot keeps the spread, the skew, and the stray values that an average quietly erases.

Comparing Multiple Groups with Box Plots

Lining several groups up next to one another is probably the single most frequent reason people reach for a box and whisker diagram.

Side-by-Side Layout

Anchor every group to one common axis, then read across them:

• Center: Are the medians sitting at noticeably different heights?
• Spread: Is one group far more scattered than the rest?
• Shape: Do the distributions look balanced, or do they lean to one side?
• Outliers: Which groups are throwing off stray points?

Two Categorical Variables

If two factors organize your data at once, say treatment arm crossed with sampling time, grouped or faceted layouts keep both dimensions readable:

# Grouped box plot in Seaborn
sns.boxplot(x='Timepoint', y='Value', hue='Treatment', data=df)

# Faceted box plot in ggplot2
ggplot(df, aes(x = Treatment, y = Value)) +
  geom_boxplot() +
  facet_wrap(~Timepoint)

Interpreting What You See

Common Mistakes to Avoid When Drawing Box Plots

A handful of habits crop up again and again, and each one can quietly mislead a reader or chip away at the trust your figure deserves.

Mistake 1: Applying Box Plots to Very Small Samples

Problem: Feed a box plot five observations and it still dutifully renders a box and two whiskers, even though every feature now hangs on one or two numbers. The picture looks far more settled than the evidence warrants.

Fix: Below roughly ten observations, switch to a strip plot or dot plot that puts each value on the page individually.

Mistake 2: Overlooking Bimodal Distributions

Problem: When a dataset hides two separate clusters, the resulting box can look perfectly ordinary while erasing the twin peaks beneath it.

Fix: Layer a strip plot over the box (geom_jitter in R, sns.stripplot in Python), or reach for a violin plot that draws the whole shape.

Mistake 3: Inconsistent Whisker Definitions

Problem: Whisker rules vary by tool. One stretches to the extremes, another halts at 1.5 x IQR, and a third pins the ends to fixed percentiles such as the 5th and 95th.

Fix: Spell out the rule you used right in the caption. Tukey's 1.5 x IQR convention is the version most readers will assume.

Mistake 4: Missing Axis Labels

Problem: Strip away the axis labels, units, and scale, and your readers are left guessing what any of the numbers actually mean.

Fix: Label every axis clearly and name the units. Hold the scale constant across groups so the comparison stays honest.

Mistake 5: Crowding Too Many Groups Into One Panel

Problem: Cram a dozen or more boxes onto one panel and precise reading goes out the window.

Fix: Sort the groups by median, break them across faceted panels, or bundle related categories. Seven boxes per panel is a sensible ceiling.

Mistake 6: Choosing Inaccessible Colors

Problem: Hues that blur together, or that colorblind readers cannot tell apart, add visual noise without carrying any information.

Fix: Reach for a scientific color palette designed with accessibility in mind, and let color mark real group differences rather than decorate.

Advanced Variations on the Standard Box Plot

The classic box and whisker diagram has spawned a family of richer formats, each adding a layer of analytical detail.

Violin Plots

A violin plot drapes a kernel density estimate over the box plot skeleton so the entire distribution becomes visible. It pays off most when the data is multimodal, exactly the case a plain box plot would gloss over.

Notched Box Plots

Cutting notches into the box near the median sketches an approximate 95% confidence interval. Two boxes whose notches stay apart point to medians that likely differ, no formal test required.

Letter-Value Plots

Letter-value plots, introduced by Heike Hofmann and collaborators, push the quartile idea outward to further quantile levels. They shine on large datasets, where the bare five-number summary leaves real detail in the tails unaccounted for.

Raincloud Plots

A raincloud stacks three views in one frame: a half-violin, a box plot, and a jittered strip of raw points. Now common in psychology and neuroscience work, it keeps the individual data honest while still offering a distributional summary.

Box Plots with Jittered Points

Scattering the actual observations across the box as a jittered strip has become standard, and sometimes mandatory, journal practice. The hybrid keeps both the summary statistics and the raw evidence in plain sight.

Blending chart types is increasingly the norm in current research. A time series, for instance, can carry box plots at each time point to expose how the distribution shifts along the way.

Box Plot Interpretation Quick Reference

Frequently Asked Questions

Summary

More than fifty years on, the box and whisker diagram still earns its keep by squeezing a surprising volume of information into very little ink. Whether you are a student meeting quartiles for the first time, a researcher pitting treatment arms against each other, or an analyst combing through a large dataset, knowing how to build and read one pays off.

To recap the workflow, drawing a box and whisker diagram comes down to:

1. Order the data from lowest to highest
2. Work out the five-number summary: minimum, Q1, median, Q3, maximum
3. Find the IQR and use it to set the fences that flag outliers
4. Lay down the box from Q1 to Q3 with the median marked inside
5. Run the whiskers to the most extreme readings that still sit within the fences
6. Mark each outlier as its own point past the whisker ends
7. Finish with labels, a title, and units so the chart explains itself

For more on assembling figures that actually communicate, browse our posts on data visualization best practices, designing scientific diagrams, and creating publication-quality figures.

Additional Resources

• Box Plot - Wikipedia
• A Complete Guide to Box Plots - Atlassian
• Box Plot Explained with Examples - Statistics By Jim
• Reading a Box and Whisker Plot - Simply Psychology
• Research Data Visualization Best Practices
• Scientific Color Palette Guide
• Best Fonts for Scientific Figures

Want polished box plots without touching a line of code? Try Figviz and turn your raw numbers straight into publication-ready charts.