# Beeswarm Plume: 9 use cases to read a distribution well

> Step by step: animate dots from one year to the next, compare with last year, raincloud, box plot, in-visual search, large volumes and group order.

- Language: English
- Web page: https://plumevisuals.com/en/tutorials/beeswarm-plume-cas-pratiques
- Published: 2026-10-21
- Visual: Beeswarm Plume, technical reference: https://plumevisuals.com/ai/visuals/beeswarm.md
- If an option name in this tutorial differs from the technical reference, the reference is right (it is generated from the visual's code).

These recipes start from a Beeswarm Plume already in place. If you do not have one yet, start with the [getting started guide](/en/tutorials/beeswarm-plume-prise-en-main): the examples reuse its data (`Stores` table, `Total sales` measure). Features marked ★ are premium: they work without a license, with a watermark.

## How do I animate dots from one year to the next ★?

You have one value per item and per year (price per m² by town, sales by store and year) and you want to see the swarm move.

1. Drop the year column (or a date on January 1) into **Play axis ★**. Use the column itself, not the date hierarchy.
2. A play bar appears under the swarm: click ▶. Each dot **glides** from its value in one year to its value in the next; medians glide with it; the current year is written in large type.
3. Set the speed in **Format › Playback ★ › Playback › Duration of a frame (ms)**, or with the ×1 button of the bar (×0.5 to ×4). Turn on **Loop** for a presentation.
4. Click an item: it stays selected in every year, so you can follow its path.

Result: dots never overlap, even while moving, and keep the same size from one year to the next. The play axis is a ★ feature: without a key, the "Premium preview" watermark appears as soon as the field holds data. With **Format › Playback ★ › More settings › Dot history** = *Selected dots*, a trail shows the past values of the item you follow.

> **Note:** A town that exists only in some years fades in when it appears and fades out when it leaves. "Reduce motion" (Format › Style › Animations, or Windows) turns transitions off.

## How do I compare each dot with last year ★?

You want to see who moves up and who moves down, not only where each one stands.

1. Create the reference measure:

```dax
Sales PY = CALCULATE ( [Total sales], SAMEPERIODLASTYEAR ( 'Date'[Date] ) )   -- or a "Sales 2024" column
```

2. Drop it into **Compare with ★**. Each dot gets a stem from last year's value and a hollow dot at that value.
3. In **Format › Comparison ★**, turn on **Color dots by their gap**: green if favorable, red if unfavorable.
4. If a decrease is good news (lead times, costs), turn off **An increase is favorable**.

Result: each dot's tooltip gives the reference and the gap, as a value and in %, with ▲ or ▼ written (the direction never relies on color alone).

## How do I build a raincloud plot ★?

A raincloud plot combines three views of a distribution: the density, the box and the dots.

1. **Format › Swarm › Layout › Side** = *One side (above or right)*.
2. **Format › Statistics › Violin ★**: turn on **Violin (density)**.
3. **Format › Statistics › Box plot ★**: turn on **Box plot**.

Result: in each band, the dots on one side, half a violin on the other, a thin box in between. Values beyond the whiskers get a red ring.

## How do I show a box plot behind the dots ★?

1. **Format › Statistics › Box plot ★**: turn on **Box plot**.
2. **Whiskers**: *1.5 × interquartile range (Tukey)* (the standard), *Minimum and maximum*, *5th and 95th percentiles* or *10th and 90th percentiles*.
3. **Ring the outliers**: dots beyond the whiskers get a ring.
4. To name them: **Format › Data labels › Labeled dots** = *Outliers (1.5 × interquartile range) ★*.

Quartiles are computed like `PERCENTILE.INC` in DAX: a measure gives the same values.

## How do I let readers search for an item ★?

In a swarm of 500 dots, a reader wants to find "their" store.

1. **Format › Search ★ › Search box in the visual ★**: turn it on. For a video or a presentation, **Hide while playing (play axis)** (on by default) and **Hide in reading view** remove the box.
2. In reading view, type a few letters in the box at the top of the visual: dots whose name contains that text stand out (case and accents ignored) and are labeled; the number found is shown.
3. Press Enter: the found dots are selected and filter the page. Esc clears the search.

To highlight items chosen in advance, type their names in **Highlight ★ › Highlighted dots ★** (one per line), or drop a measure into **Highlight ★** (not blank and not 0 = highlighted).

## How do I show 100,000 dots?

1. Drop the detailed field into **Details**: Power BI sends 30,000 rows, then the visual asks for the rest, up to 200,000 rows.
2. Keep **Shrink dots so that everything fits** on: dots get very small (never below 0.6 px) but none spills out of its band.
3. Enlarge the visual and keep the number of groups low so that dots stay visible.

Beyond about 20,000 dots in one band, the visual switches to a fast layout (staggered grid): each dot moves at most one radius away from its exact value.

## How do I sort groups by median ★?

1. **Format › Swarm › Groups (bands) › Group order** = *Median, descending ★* (or *Median, ascending ★*).
2. Groups do not move during playback: the median used is the one of all years.

Other options, free: *Number of dots*, *Name (A → Z)*, or the data order (the sort of the column in the model, for instance an "Order" column).

## How do I build a dot histogram ★?

1. **Format › Swarm › Layout › Layout** = *Sieve (dot columns) ★*.
2. **Side** = *One side (above or right)* for columns standing on a baseline.

Result: each column holds the dots of an interval one dot wide, stacked in value order: a histogram where you can still hover and select each item.

## How do I title the swarm with the current median?

Dynamic formatting drives any setting with a measure (free).

1. Create the measure:

```dax
Swarm title =
"Sales of " & COUNTROWS ( ALLSELECTED ( Stores ) ) & " stores · median "
    & FORMAT ( CALCULATE ( MEDIAN ( Stores[Sales] ), ALLSELECTED ( Stores ) ), "$#,0" )
```

2. Drop it into **Dynamic format** and rename it for this visual `title.text` (right-click the field › Rename for this visual).

Result: the title follows the page slicers. The list of keys is in [dynamic formatting](/en/tutorials/formats-dynamiques).

## How do I keep the same scale when filtering?

A `Format` slicer removes the stores of the other format: the value axis refits to the remaining dots, and the eye can no longer compare. Pin the bounds with measures that ignore this filter.

1. Create the measures:

```dax
Axis min =
CALCULATE ( MIN ( Stores[Sales] ), REMOVEFILTERS ( Stores[Format] ) )  -- the lowest sales, all formats

Axis max =
CALCULATE ( MAX ( Stores[Sales] ), REMOVEFILTERS ( Stores[Format] ) )
```

2. Drop them into **Dynamic format**, then rename them for this visual `valAxis.min` and `valAxis.max`.
3. Pick a format in the slicer: the axis no longer moves, only the dots change.

Don't use `ALLSELECTED ()`: it follows slicers, so the scale would still move. `REMOVEFILTERS ( Stores )` (or `ALL ( Stores )`) ignores every filter. A highlight sent by another visual does not change the scale (**Interactions › Scale during cross-highlighting**).