Make a scatter chart of thousands of points readable with hexagons

In short

What is a hexbin? When a scatter chart holds thousands of points, they pile up on top of each other: you see a blob, not where the points concentrate. A hexbin cuts the chart into small hexagonal bins (like a honeycomb) and colors each bin by the number of points inside: pale when there are few, dark when there are many.

Mini example. Out of 15,000 homes sold, placed by area (X axis) and price (Y axis), the bin "55 to 65 m², £150,000 to £180,000" holds 330 homes: it is dark red; the bin "180 m², £1 million" holds 2: it is pale yellow.

• What the visual does. Hexbin Plume groups up to 300,000 points into bins, without sampling (no point is hidden). The color follows the number of points or a calculation on a measure (sum, average, median, minimum, maximum). Clicking a bin filters the page on all its points; when you zoom in, the bins get finer, then the points appear. In premium (★), the density changes from one year to the next with a play axis, and the visual adds advanced color scales (log, diverging), compares two groups, shows the change between two periods, draws proportional bins, contour lines, marginal histograms, a trend line and small multiples.
• When to use it. Lots of points: customers (average basket and frequency), orders (amount and lead time), sensors (temperature and humidity), homes (area and price), employees (seniority and salary).
• When to prefer something else. A few dozen points you want to name one by one: Scatter Plume. Two categories in rows and columns: Matrix Plume with a heat map.
• What you will learn. Prepare a one-point-per-row table and two measures, insert the visual, read the density and the tooltip of a bin, filter the page by clicking a bin, zoom in.
• What you need: Power BI Desktop, the Hexbin Plume visual installed (see the shared basics), a table with one ID per row and two numeric columns. The sample report PlumeHexbin-Sample-EN, in your "Downloads" area, shows 15,000 fictitious homes (page 1), a value per bin, the dominant city and the play axis (page 2) and the ★ features (page 3).

The data to provide

The shape. One row per point: an ID (the home, the customer, the order) and two numbers (the position on X and on Y). A category, a period and a third measure are optional.

Home	City	Year	Area	Price	Rent
L00001	Bristol	2023	48	212,000	760
L00002	York	2021	95	218,000	1,090
L00003	Leeds	2024	62	118,000	610
L00004	Bristol	2022	71	305,000	1,100
L00005	Norwich	2019	120	214,000	1,200
L00006	York	2024	55	142,000	650
L00007	Leeds	2020	38	66,000	380
L00008	Bristol	2024	33	160,000	520

To try without a file: Modeling › New table, then paste:

Homes =
DATATABLE (
    "Home", STRING, "City", STRING, "Year", INTEGER, "Area", INTEGER, "Price", INTEGER, "Rent", INTEGER,
    {
        { "L00001", "Bristol", 2023, 48, 212000, 760 },
        { "L00002", "York", 2021, 95, 218000, 1090 },
        { "L00003", "Leeds", 2024, 62, 118000, 610 },
        { "L00004", "Bristol", 2022, 71, 305000, 1100 },
        { "L00005", "Norwich", 2019, 120, 214000, 1200 },
        { "L00006", "York", 2024, 55, 142000, 650 },
        { "L00007", "Leeds", 2020, 38, 66000, 380 },
        { "L00008", "Bristol", 2024, 33, 160000, 520 }
    }
)

Eight rows give eight bins: enough to follow the steps. To see a real hexbin, use the sample report (15,000 homes) or your own table.

The measures. The X and Y axes expect measures (a calculation Power BI redoes for the filters, written in DAX; see the glossary). With the Details field filled with the ID, each measure is computed for a single home: an average returns its value.

Average area = AVERAGE ( Homes[Area] )   -- one home per point: the average is its area
Average price = AVERAGE ( Homes[Price] )
Average rent = AVERAGE ( Homes[Rent] )

Pitfalls.

• Fill Details with the ID: without it, Power BI adds everything up and the visual gets a single point.
• If your table has several rows per home (several sales), the ID is still the right grain; pick the measure (sum, average) that makes sense.
• Rows where X or Y is blank are ignored. With a log scale, zero or negative values too (the visual says so in edit mode).

Step by step: your first hexbin in 10 minutes

1. In the Visualizations pane, click the Hexbin Plume icon. An empty frame appears, with a start screen listing the fields to drop; enlarge it.
2. Drag the Home column into Details (one point each).
3. Drag Average area into X axis, then Average price into Y axis. The bins appear, with the title "Average price vs Average area" and the legend "Number of points".
4. Read the legend: from the palest bin (1 point) to the darkest (the maximum). Dark bins show where homes concentrate.
5. Hover a bin: it is outlined, its points appear, and the tooltip gives the number of homes (and their share), the area and price ranges of the bin.
6. Click the darkest bin: the other visuals of the page filter on its homes (cross-filtering, see the glossary) and the other bins fade. Click the bin again to cancel.
7. Drag the mouse over the chart: a rectangle selects every home inside.
8. Hold Shift and drag over the small areas: the chart zooms in, the bins are recomputed finer. Double-click to go back to the full view.
9. Open Format › Bins › Shape and size and try Bin size = Radius (px) with 20: fewer, larger bins.
10. Drag City into Legend (category), then Format › Data colors › Coloring = Dominant category (Legend field): each bin takes the color of its most frequent city.

Expected result: a readable density map, a detailed tooltip per bin, a click that filters the page, a zoom that refines the bins.

Troubleshooting

What you see	Likely cause	Solution
A single bin	The Details field is empty: Power BI adds everything up	Drag the ID into Details
"No point to show"	X or Y holds no numbers	Check the measures (numeric type)
Every bin has the same color	One point per bin (little data or bins too small)	Format › Bins › Bin size: larger radius
The bins are too big	Automatic size with few points	Radius in px or number of columns
Coloring by city does not apply, message in edit mode	The Legend field is empty	Drag a category into Legend
Points disappear with a log axis	Zero or negative values	Linear scale, or exclude those rows
"Premium preview" watermark	A ★ feature is used without a key	Key in Format › Plume Studio License, or a free setting
Clicking a bin is slow	A dense bin holds tens of thousands of points	Zoom in first, or filter the page

Key takeaways

• One row per point and the ID in Details; X and Y as measures.
• Every point is counted, up to 300,000: no sampling.
• Clicking a bin filters the page on all its points; Shift + drag zooms in and refines the bins.
• The dominant category, the linear and square root scales and the zoom are free; the play axis, the log and diverging scales, comparing two groups, the change between frames and the ★ analytics work with a watermark until you enter a key.

Learn more

• Hexbin Plume how-to guides: recipes for common needs.
• Hexbin Plume fields and settings: every field well and every setting.
• Getting started with Plume visuals: installation, license, styles and shared settings.
• The visual's page: Hexbin Plume. A question: support.