When should I use a hexbin plot?

Very dense scatter data (1000+ points) ● Overplotting is a problem ● Density estimation

When should I avoid a hexbin plot?

Sparse data ● When individual points matter

What are common mistakes with hexbin plots?

Bin size too large hides patterns ● Using rainbow colormap

Relationship

Hexbin Plot

Aggregates dense scatter points into hexagonal bins colored by count.

NYC Pickup Density

Taxi pickup locations (Manhattan)

View data (2,050 rows)

Chart data table: NYC Pickup Density
Longitude	Latitude
-73.98	40.76
-73.98	40.76
-73.99	40.75
-73.98	40.76
-73.98	40.75
-73.99	40.76
-73.98	40.76
-73.98	40.76
-73.98	40.77
-73.98	40.75
-74.00	40.76
-73.99	40.76
-73.98	40.76
-73.98	40.75
-73.99	40.75
-73.97	40.75
-73.98	40.76
-73.98	40.76
-73.98	40.75
-73.98	40.76
-73.98	40.76
-73.98	40.76
-73.99	40.75
-73.97	40.76
-73.98	40.75
-73.99	40.76
-73.98	40.76
-73.99	40.76
-73.99	40.76
-73.98	40.75
-73.98	40.75
-73.99	40.76
-73.99	40.76
-73.99	40.76
-74.00	40.76
-73.97	40.76
-73.99	40.76
-73.98	40.76
-73.99	40.77
-73.99	40.76
-73.98	40.76
-74.00	40.76
-73.99	40.76
-73.99	40.76
-73.98	40.75
-73.98	40.76
-73.99	40.77
-73.98	40.75
-73.98	40.77
-73.98	40.75
-73.99	40.76
-73.98	40.76
-73.99	40.76
-73.99	40.76
-73.99	40.75
-73.99	40.76
-73.98	40.76
-73.99	40.75
-73.99	40.76
-73.99	40.76
-73.99	40.76
-73.99	40.76
-73.98	40.77
-73.98	40.76
-73.99	40.75
-74.00	40.77
-73.98	40.77
-73.98	40.76
-74.00	40.76
-73.98	40.76
-73.98	40.76
-73.99	40.75
-73.99	40.76
-73.98	40.75
-73.99	40.75
-73.99	40.76
-73.99	40.77
-73.99	40.77
-73.99	40.77
-73.98	40.75
-73.98	40.77
-73.98	40.76
-73.98	40.75
-73.99	40.75
-73.98	40.75
-73.99	40.75
-73.97	40.76
-74.00	40.75
-73.99	40.76
-73.99	40.75
-73.98	40.76
-73.99	40.75
-73.99	40.75
-73.98	40.76
-73.98	40.75
-73.99	40.75
-73.98	40.75
-73.97	40.76
-73.98	40.75
-73.99	40.76
…and 1,950 more rows. Download CSV for the full dataset.

Make a hexbin plot with your data

Use a hexbin plot when…

Very dense scatter data (1000+ points)
Overplotting is a problem
Density estimation

Avoid when…

Sparse data
When individual points matter

Data it needs

Property	Value
Min Rows	20
Min Columns	2
Column Types	numbernumber

Visual anatomy

Marks

hexagon

Channels

positioncolor-luminance

Axes

x-quantitativey-quantitative

Guiding principles

ColorPerceptually Uniform Colormaps
Use sequential colormap for bin counts
IntegrityZero Baselines
Choose bin size where the median bin holds roughly 5-20 points — too large erases pattern, too small recreates the overplotting problem

Consider instead

Scatter Plot

For sparse data

2D Contour Plot

For smooth density estimation

Common mistakes

Bin size too large hides patterns
Using rainbow colormap

History

Proposed by Carr et al. (1987) as an improvement over binned scatter plots.

Accessibility notes

Provide density summary in text. Use perceptually uniform colormap.

Got data? Let's see what works.

Drop your CSV. You'll get a Hexbin Plot plus four alternatives - ranked by which one actually fits your data best.

Try it free