Sector Chart Assertions¶

Status: draft Scope: pie and donut charts Source: Chart Design Notes, especially Notes 31-41

This document is the first assertion-family prototype for turning design notes into implementation-oriented design decisions.

Batch Synthesis¶

The sector-chart notes point toward a coherent posture rather than a loose set of independent tips.

The implied design stance is:

treat pie and donut as one chart family built around coarse part-to-whole reading rather than precision-by-geometry
prefer restrained, highly legible sector charts over decorative or over-segmented ones
keep orientation, labeling, and the definition of the whole inside the chart as much as possible
use donut-specific affordances only when they do real work, especially for clarifying the whole
keep circular ordering, rotation, palette, and labeling under explicit control rather than treating them as incidental defaults

The main tensions inside the batch are:

icon vs container
pies have stronger solid presence
donuts offer calmer composition and a potential annotation surface
geometry vs text
wedges provide coarse share structure
labels often need to carry exactness, identity, and orientation
simplicity vs scaffolding
sector charts benefit from clean silhouettes
but sometimes need explicit aids such as value labels, separators, or rim ticks
local optimization vs system consistency
a single pie may want one ordering or seam
repeated pies may need stricter consistency for comparison

If these assertions were handed to an AI to propose theme + structure, the result should likely favor:

a conservative sector-chart palette
direct or near-mark labeling before detached legends
explicit control of ordering and rotation
optional thin separators
optional but default-off sector scaffolding
strong support for total/whole orientation, especially in donut centers

Assertions¶

Assertion 1¶

Title: Pie And Donut Are One Chart Family
Source Notes: Note 31
Assertion: Pie and donut charts should be treated as one chart family with the same primary quantitative channel: sector angle.
Why It Matters: This keeps pie-vs-donut decisions grounded in compositional role rather than pretending they are fundamentally different encodings.
Applies To: pie, donut
Claim Type: chart-family definition
Implementation Home: docs / chart recommendation
Vega-Lite Fit: native
Dataface Path: treat pie and donut as arc-family variants differentiated mainly by inner radius and annotation opportunities.
Decision Strength: dependable default
Feasibility: easy
Cost: low
Theme / Structure Consequence: shared family-level guidance; donut-specific structure variants only where needed
Evaluation Check: docs and recommendation logic describe donut as a modified pie rather than a separate quantitative family
Open Questions: none

Assertion 2¶

Title: Wedge Order Follows Reading Task
Source Notes: Note 32
Assertion: Wedge sort order should be chosen by reading task, not convention alone.
Why It Matters: Different tasks favor different circular sequences: rank reading, semantic sequence, and cross-chart comparison do not always want the same order.
Applies To: pie, donut
Claim Type: default / exception
Implementation Home: docs / renderer / authoring controls
Vega-Lite Fit: native
Dataface Path: support descending order, input order, and semantic/custom order as first-class options.
Decision Strength: context-sensitive adaptation
Feasibility: easy
Cost: low
Theme / Structure Consequence: no theme effect; structure/renderer must expose ordering policy clearly
Evaluation Check: author can produce rank-first, semantic-order, and stable-comparison variants without hacks
Open Questions: how best to express semantic/custom order in authored config

Assertion 3¶

Title: Single Sector Charts Default To A Noon-Anchored Reading Frame
Source Notes: Note 32, Note 37
Assertion: Single pie and donut charts should default to a noon-anchored reading frame with an order chosen for the reader task.
Why It Matters: Circular charts lack a natural start point; noon gives readers a familiar seam and stable entry.
Applies To: pie, donut
Claim Type: opinionated default
Implementation Home: structure / renderer
Vega-Lite Fit: native
Dataface Path: default start angle to a noon-oriented seam while keeping start angle overridable.
Decision Strength: opinionated default
Feasibility: easy
Cost: low
Theme / Structure Consequence: structure default for arc start angle and directional policy
Evaluation Check: default pies and donuts feel upright and legible; override remains possible
Open Questions: exact start-angle convention to use relative to Vega-Lite arc orientation

Assertion 4¶

Title: Rim Ticks Are Optional Sector Scaffolding
Source Notes: Note 33
Assertion: Rim ticks on pie and donut charts should be available as optional quantitative scaffolding, but default off.
Why It Matters: Sector charts are weak at precise share reading; light outer ticks can restore measurement aids when needed.
Applies To: pie, donut
Claim Type: exception / renderer feature
Implementation Home: renderer / annotation system
Vega-Lite Fit: unsupported or awkward
Dataface Path: custom arc-adjacent annotation layer with preset divisions such as quarters, tenths, or clock-face.
Decision Strength: authored specificity
Feasibility: moderate
Cost: medium
Theme / Structure Consequence: default off; optional structure preset if implemented
Evaluation Check: ticks can be enabled without overpowering wedges or labels
Open Questions: whether to support as a true feature now or hold as later renderer enhancement

Assertion 5¶

Title: Sector Charts Prefer Local Geometry-Aware Labeling
Source Notes: Note 34
Assertion: Sector charts should prefer local, geometry-aware labeling over detached legends.
Why It Matters: Circular forms impose high lookup cost; labels should keep readers in the chart whenever possible.
Applies To: pie, donut
Claim Type: opinionated default
Implementation Home: docs / renderer / annotation logic
Vega-Lite Fit: partial
Dataface Path: prefer direct labels when wedge geometry supports them; allow wedge-face, arc-adjacent, or outside labels depending on slice size.
Decision Strength: opinionated default
Feasibility: moderate
Cost: medium
Theme / Structure Consequence: structure should prefer direct label modes before legend fallback
Evaluation Check: charts with few readable slices do not default to detached legend-only identification
Open Questions: how much geometry-aware placement can be done natively versus custom

Assertion 6¶

Title: Sector Legends Must Mirror Wedge Order
Source Notes: Note 34, Note 32
Assertion: If a legend is used on a sector chart, its order should match wedge order.
Why It Matters: Legends are already costly on circular charts; mismatched order creates a second competing sequence.
Applies To: pie, donut, legend
Claim Type: renderer contract / default
Implementation Home: renderer / legend defaults
Vega-Lite Fit: likely native or easy
Dataface Path: synchronize legend ordering with arc ordering whenever legend is shown.
Decision Strength: dependable default
Feasibility: easy
Cost: low
Theme / Structure Consequence: no theme effect; renderer/structure must keep order synchronized
Evaluation Check: legend list and wedge scan path align in default and custom sorts
Open Questions: none

Assertion 7¶

Title: Sector Charts Need Readably Large Slice Structures
Source Notes: Note 35
Assertion: Sector charts should be limited to slice structures large enough to support clear comparison and labeling.
Why It Matters: Tiny slices are where sector charts fail first: comparison weakens, labels collapse, and noise accumulates.
Applies To: pie, donut
Claim Type: chart-family suitability rule
Implementation Home: docs / validation / chart recommendation
Vega-Lite Fit: not relevant
Dataface Path: recommendation and validation should caution against over-sliced sector charts.
Decision Strength: opinionated default
Feasibility: easy
Cost: low
Theme / Structure Consequence: no direct theme effect; may drive recommendation thresholds
Evaluation Check: over-segmented sector charts trigger warnings or alternative-chart suggestions
Open Questions: what threshold signals should be used: count, minimum angle, or both

Assertion 8¶

Title: Long Tails Should Be Aggregated When Honest
Source Notes: Note 35
Assertion: Sector charts should discourage long tails by aggregating residual slices into Other when that preserves honesty.
Why It Matters: Residual fragmentation destroys the bold share structure sector charts depend on.
Applies To: pie, donut, Other
Claim Type: default / exception
Implementation Home: query guidance / docs / recommendation logic
Vega-Lite Fit: requires pre-processing, not native arc config
Dataface Path: encourage upstream aggregation or authoring assistance for residual categories.
Decision Strength: context-sensitive adaptation
Feasibility: moderate
Cost: medium
Theme / Structure Consequence: none directly; affects chart-preparation guidance
Evaluation Check: charts with long tails are either aggregated or redirected to a better form
Open Questions: whether Dataface should only suggest this or eventually automate it

Assertion 9¶

Title: Sector Palettes Must Be Adjacency-Aware
Source Notes: Note 39
Assertion: Sector-chart palettes should be chosen with adjacent wedge interaction in mind, not just category distinctness in isolation.
Why It Matters: Wedges necessarily touch; boundary blur and color vibration happen locally, not abstractly.
Applies To: pie, donut, palette
Claim Type: theme rule
Implementation Home: theme / docs / palette policy
Vega-Lite Fit: native for palette application, non-native for adjacency evaluation
Dataface Path: sector-specific palette guidance and possibly sector-safe palette sets.
Decision Strength: opinionated default
Feasibility: moderate
Cost: low to medium
Theme / Structure Consequence: sector charts may need more conservative categorical palettes than other chart families
Evaluation Check: neighboring slices remain distinguishable without excessive visual noise
Open Questions: whether adjacency-aware palette validation should be explicit tooling or guidance only

Assertion 10¶

Title: Thin Separators Are Legitimate Boundary Aids
Source Notes: Note 39
Assertion: Thin separators should be available as optional boundary aids for sector charts, but should not substitute for palette restraint.
Why It Matters: When wedges must touch, a quiet boundary aid can restore clarity without making color do all the work.
Applies To: pie, donut, stroke, separator
Claim Type: default / exception
Implementation Home: theme / structure / renderer
Vega-Lite Fit: native
Dataface Path: support subtle arc stroke defaults or optional sector separators, tuned lightly.
Decision Strength: context-sensitive adaptation
Feasibility: easy
Cost: low
Theme / Structure Consequence: thin, quiet separators may be a sector-chart option, not a universal default
Evaluation Check: separators clarify boundaries without making the chart feel coarse or mechanical
Open Questions: whether separator default should vary by slice count or ring thickness

Assertion 11¶

Title: Sector Charts Need Explicit Precision Labels When Precision Matters
Source Notes: Note 40
Assertion: Sector charts should usually carry explicit values when the task requires precision beyond coarse share recognition.
Why It Matters: Sector geometry is good for broad part-to-whole structure, not fine discrimination.
Applies To: pie, donut, value labels, percent labels
Claim Type: default / exception
Implementation Home: renderer / docs / annotation defaults
Vega-Lite Fit: partial
Dataface Path: support percent labels, nominal labels, or both when slice count and space allow.
Decision Strength: context-sensitive adaptation
Feasibility: moderate
Cost: medium
Theme / Structure Consequence: sector presets may offer %, value, or % + value label modes
Evaluation Check: precision-oriented sector charts do not rely on wedge judgment alone
Open Questions: what default label hierarchy should be used when only one form fits

Assertion 12¶

Title: Sector Charts Must Explicitly Establish The Whole
Source Notes: Note 41, Note 36
Assertion: Sector charts must explicitly establish the whole being partitioned, and donut centers should preferentially be used for that when helpful.
Why It Matters: A pie or donut is only meaningful as a partition of some total; shares of what? is a first-order orientation problem.
Applies To: pie, donut, title, total, center annotation
Claim Type: dependable default
Implementation Home: docs / content model / annotation system
Vega-Lite Fit: partial
Dataface Path: support explicit whole/total framing through title, subtitle, total, or donut-center annotation.
Decision Strength: dependable default
Feasibility: moderate
Cost: medium
Theme / Structure Consequence: donut center should preferentially hold whole/total context when used; title system should distinguish domain from whole
Evaluation Check: a reader can answer shares of what? immediately from the chart framing
Open Questions: how much whole/total support belongs in Vega-Lite title/subtitle vs custom annotation

Implementation Mapping Considerations¶

This batch should not yet be treated as a one-assertion-at-a-time mapping exercise. Several assertions imply one another or only make full sense as a coherent set.

The most important batch-level implementation implications are:

sector charts likely need a chart-family-specific structure posture rather than only generic arc rendering
some of the most important sector behaviors are not purely theme decisions
label placement
whole/total orientation
legend fallback behavior
ordering and seam behavior
several high-value features are only partially native in Vega-Lite
geometry-aware local labeling
donut-center annotation as a design-system feature
rim ticks

The likely first implementation frontier is:

structure defaults
noon-anchored seam
ordering conventions
legend suppression or fallback behavior
theme defaults
conservative sector palette posture
optional thin separators
docs / recommendation behavior
slice-count caution
Other guidance
explicit-whole guidance
renderer or annotation enhancements
local label strategy
donut-center annotation support
optional rim ticks

In other words, the batch suggests a sector-chart design system that is partly:

docs and recommendation guidance
chart-family-specific structure behavior
selective renderer improvements

not just a palette tweak.

Residual Knowledge¶

The source notes contain useful knowledge that has not yet been promoted into assertions.

Important residual knowledge includes:

the more rhetorical distinction that pies are stronger icons while donuts are stronger containers
useful for framing and recommendation
not yet promoted as a separate assertion because it is evaluative and partially downstream of other choices
the historical contrast around rim ticks
strong enough to justify optional support
not strong enough to support a confident causal story about why ticks disappeared
the richer family of alternative wedge arrangements
bilateral arrangements around noon
bottom-centered dominant wedges
other seam metaphors These remain important examples and may later motivate more specific assertions or authored presets.
the detailed % vs nominal vs % + nominal labeling discussion
captured at a high level in Assertion 11
but not yet promoted into a more detailed assertion family about precision-label modes
the distinction between boundary trouble and vibration trouble in adjacent colors
partially captured in Assertion 9
still useful as explanatory vocabulary for future palette work

These residual items should not be treated as lost. They remain part of the design knowledge base and may later become assertions if implementation needs or stronger evidence justify promotion.