Visualization

When to use pie

Pie is the right choice for part-of-whole composition when the number of slices is small (typically ≤ 6) and proportional weight matters. Use the donut variant when you want to free the chart center for an inner label or a related KPI; use the rose variant for an alternative encoding where slice angle is fixed but radius reflects the value.

Use bar with chart.stack: percent instead when the categories are many, or when you want to compare composition across multiple groups. Use grid when the audience needs the actual percentages in tabular form.

Mapping

mapping.x — required. String field used as the slice name and the legend label.
mapping.y — required. Numeric field that drives the slice angle.

mapping:
  x: status
  y: order_count

Slice colors come from the platform palette in query order. To control order, sort in the underlying Malloy query.

chart shortcuts

The chart block is typed and closed.

chart.variant — "pie" (default), "donut", or "rose". Donut adds an inner radius (free center); rose switches to a rose-area layout where slice angle is uniform and radius encodes the value.
chart.inner_radius — number (pixels) or string (e.g. "55%"). Use with variant: donut; the larger the inner radius, the thinner the donut ring.
chart.outer_radius — number or string.
chart.slice_border_radius — pixels. Rounded corners on slices for a softer look.
chart.show_value_labels — boolean. Turns slice labels on. Style them with chart.label below.
chart.cross_filter — boolean, default true.
chart.height — pixel height of the viz container.

Pass-through fields (1:1 with the underlying chart):

chart.center — center of the pie. Array [x, y] of percent strings (e.g. ["50%", "50%"]) or pixel numbers.
chart.start_angle — degrees. Default 90.
chart.min_angle — degrees. Minimum angle for tiny slices to remain visible.
chart.rose_type — when using a rose layout: "radius" or "area".

Slice labels

chart.label is an object that styles slice labels when chart.show_value_labels is on:

position — "inside", "outside", "top", "bottom", "left", "right", plus the "insideLeft" / "insideRight" variants.
rotate — degrees, between -90 and 90.
color, font_size, font_weight.
formatter — string template. Use {b} for the slice name, {c} for the value, {d} for the percentage.
distance, align, vertical_align, clip.

chart:
  show_value_labels: true
  label:
    position: outside
    formatter: "{b}: {d}%"

Legend & tooltip

chart.legend controls the legend; chart.tooltip controls the hover tooltip. Both share the same shape as on bar — see that page for the full sub-block reference.

Pie-specific tip: for a donut variant with the legend on the right, set center: ["42%", "50%"] to nudge the donut left so the legend has more room.

format

format.y or format[<value_field_name>] — pattern for slice tooltip values and inside-slice labels.
format at the root — fallback.

Cross-filter behavior

Clicking a slice cross-filters the rest of the dashboard by the slice label.
The clicked field (the slice-label field, e.g. status) must be declared as a parameter in at least one model used by the dashboard, or the click is silently ignored.
Disable per viz with chart.cross_filter: false.

See Cross-filtering for the full mechanism.

Worked examples

Donut with outside labels and bottom legend:

id: ec_orders_by_status_pie
title: Orders by Status
query: "models/ec_revenue.malloy::by_status"
type: pie
mapping:
  label: status
  value: order_count
chart:
  variant: donut
  inner_radius: "55%"
  outer_radius: "78%"
  show_value_labels: true
  label:
    position: outside
    formatter: "{b}: {d}%"
  legend:
    show: true
    position: bottom
format:
  order_count: "#,##0"
published: true

Solid pie with inside labels:

type: pie
mapping:
  label: category
  value: revenue
chart:
  show_value_labels: true
  label:
    position: inside
    color: "#fff"
    font_weight: bold
  legend:
    show: false
format:
  revenue: "$#,##0"

Rose layout (slice angle uniform, radius encodes value):

chart:
  variant: rose
  rose_type: area
  start_angle: 0
  show_value_labels: true

Common pitfalls

Too many slices. A pie with more than ~6 slices becomes hard to read. Sort + limit in the Malloy query, or aggregate small slices into an "Other" bucket.
Slices with zero or near-zero values disappear. Use chart.min_angle to enforce a minimum visible angle, or filter zeros out in the query.
Labels overlap on a tight chart. Move them inside (chart.label.position: "inside") or rotate.
Donut inner content collides with labels. When you put a KPI in the donut center via dashboard layout, set chart.label.position: "outside" so the slice labels do not collide.
Slice colors are not what you expected. Colors are assigned in query-result order from the platform palette. Sort the query so the largest or most important slice gets the dominant color.
Cross-filter clicks have no effect. The slice-label field must be declared as a parameter in at least one model used by the dashboard.

Visualization — pie