Dot-Boxplot
A box plot with individual data points overlaid — showing both the statistical summary and the raw data that produced it.
// 01 — The chart
What it looks like
A horizontal dot-boxplot comparing response times across three server regions. Each dot represents an individual request.
// 02 — Definition
What is a dot-boxplot?
A dot-boxplot is a hybrid visualization that layers individual data points on top of a traditional box plot. The box plot provides the five-number summary — minimum, first quartile (Q1), median, third quartile (Q3), and maximum — while the overlaid dots reveal the actual distribution, clusters, gaps, and outliers in the raw data.
Standard box plots are powerful but opaque: they compress an entire dataset into five statistics, hiding important details like bimodality, data density, and gaps. By adding dots, you get the best of both worlds — the concise summary of a box plot plus the transparency of showing every data point.
Dot-boxplots are especially popular in scientific publications and statistical analysis where transparency about the underlying data is paramount. They’re sometimes called “box-and-dot plots” or “box plots with jittered points.”
Key insight: The American Statistical Association recommends showing raw data points alongside summaries. A dot-boxplot does exactly this — ensuring no details are hidden behind aggregate statistics.
// 03 — Anatomy
Parts of a dot-boxplot
// 04 — Usage
When to use it — and when not to
- You want to show both summary statistics and raw data transparency
- Your dataset is small to medium (under ~200 points per group)
- You suspect the data may be bimodal or have gaps hidden by the box
- Comparing distributions across 2–8 groups side by side
- Publishing scientific results where data transparency is required
- You need to spot outliers in context of the full dataset
- Your dataset has thousands of points — dots will overplot into a solid blob
- You have more than 10 groups — the chart becomes too wide or dense
- A violin or density plot would better reveal the distribution shape
- Your audience only needs the summary — a plain box plot may suffice
- Data is categorical or ordinal — dots imply continuous measurement
- The chart will be printed very small — dots become unreadable
// 05 — Reading guide
How to read a dot-boxplot
Follow these steps to extract maximum insight from a dot-boxplot.
Read the box first
Identify the median (center line), Q1 (left/bottom edge), and Q3 (right/top edge). The box contains the middle 50% of data — the interquartile range (IQR).
Check the whiskers
Whiskers extend to the most extreme data points within 1.5× IQR from the box edges. They show the range of 'typical' values.
Look at the dots
Do the dots cluster around the median, or are there clumps elsewhere? Gaps or multimodal clusters the box alone would hide become visible.
Spot outliers
Points beyond the whiskers are outliers. Check whether they are isolated extremes or form a secondary cluster — a crucial difference.
Compare across groups
When multiple dot-boxplots sit side by side, compare medians, spreads, and the dot patterns. One group may look similar in summary but differ dramatically in raw data shape.
// 06 — Data format
What your data should look like
A dot-boxplot needs one categorical column (the group) and one continuous column (the measurement).
| group | value |
|---|---|
| US-East | 142 |
| US-East | 198 |
| US-East | 310 |
| EU-West | 225 |
| EU-West | 189 |
| AP-South | 340 |
Code sketch — Python
import seaborn as sns
sns.boxplot(data=df, x="group", y="value")
sns.stripplot(data=df, x="group", y="value",
jitter=True, alpha=0.5, color="black")// 07 — Construction
How to build one, step by step
Compute the five-number summary for each group: min, Q1, median, Q3, max.
Draw the box from Q1 to Q3 with a vertical line at the median.
Draw whiskers from the box edges to the most extreme non-outlier points (within 1.5× IQR).
Overlay each individual data point as a dot, applying slight vertical jitter to reduce overplotting.
Distinguish outliers (points beyond whiskers) with a different marker style or emphasis.
Ensure dot transparency or small size so the box plot remains readable beneath.
// 08 — Common mistakes
Mistakes to avoid
Too many dots without jitter
When points overlap, apply random jitter along the perpendicular axis so every data point is visible.
Opaque dots hiding the box
Use semi-transparent dots (alpha 0.3–0.6) or small dot sizes so the box plot summary remains clearly readable.
Using dot-boxplots for huge datasets
With thousands of points per group, dots become a solid blob. Switch to a violin or density plot instead.
Inconsistent jitter width across groups
Keep jitter width uniform so visual density is comparable between groups.
// 09 — In the wild
Real-world examples
Biomedical research
Clinical trial papers routinely overlay patient-level outcomes on box plots to satisfy journal transparency guidelines.
Software engineering
API latency dashboards show percentile boxes with individual request dots to reveal bimodal response patterns (cache hit vs. miss).
Education
Student test scores plotted as dot-boxplots let teachers see both the class summary and each student’s position.
// 10 — At a glance
Quick reference
Category
Distribution
Data type
Continuous + categorical
Best for
Small–medium n
Max groups
~8–10
Difficulty
Intermediate
Also called
Box-and-dot plot
// 11 — Accessibility
Accessibility notes
Use distinct marker shapes (not just color) when comparing groups on the same plot
Provide an aria-label summarizing the five-number statistics and number of observations
Include a data table alternative showing percentiles and individual values
Ensure dots have sufficient contrast against the box fill — use outlined dots on light fills
Add a text description noting any visible clusters, gaps, or outlier patterns
// 12 — Variations
Variations
Sina plot
Dots are jittered proportionally to the local density — wider jitter where data is densest, forming a violin-like shape.
Beeswarm-box hybrid
Uses force-directed placement instead of random jitter so dots never overlap, combined with a box plot.
Letter-value plot + dots
Replaces the box with nested letter-value quantiles for larger datasets, overlaid with dots.
Notched dot-boxplot
Adds confidence interval notches to the median line — if notches of two groups don’t overlap, medians likely differ.
// 13 — FAQs
Frequently asked questions
What is a dot-boxplot?+
A dot-boxplot is a hybrid visualization that layers individual data points on top of a traditional box plot. The box plot provides the five-number summary — minimum, first quartile (Q1), median, third quartile (Q3), and maximum — while the overlaid dots reveal the actual distribution, clusters, gaps, and outliers in the raw data.
When should you use a dot-boxplot?+
Use a dot-boxplot when you want to show both summary statistics and raw data transparency. It also works well when your dataset is small to medium (under ~200 points per group), and when you suspect the data may be bimodal or have gaps hidden by the box.
When should you avoid a dot-boxplot?+
Avoid a dot-boxplot when your dataset has thousands of points — dots will overplot into a solid blob. It is also a poor fit when you have more than 10 groups — the chart becomes too wide or dense, or when a violin or density plot would better reveal the distribution shape.
What data do you need to make a dot-boxplot?+
A dot-boxplot needs one categorical column (the group) and one continuous column (the measurement).
What size of dataset works best for a dot-boxplot?+
Dot-Boxplot works best for Small–medium n. Outside that range the chart either looks empty or becomes too cluttered to read clearly.
Are dot-boxplots accessible to screen readers?+
Yes — a dot-boxplot can be made accessible to screen readers by pairing it with a clear text summary of the key insight, ensuring color choices meet WCAG contrast guidelines, adding descriptive alt text or aria-label to the SVG, and offering the underlying data as an HTML table fallback for assistive technologies.