Cookbook
Complete, reusable chart documents organized by the question they answer.
These are the same examples returned by the MCP server's get_examples tool. Each uses illustrative data and contains one complete document for validation, rendering, or saving. Expand an example to inspect its chart and editable definition.
Start with a simple comparison, trend, composition, or relationship. The focused recipes demonstrate data modes, ranges, uncertainty, references, and family-specific annotations. The activity-by-platform example deliberately contains a null observation and returns a warning for the visible gap.
Comparison
Quarterly revenue
Long rows, native value labels, and no redundant value axis.
Team headcount
Rank categories by value; horizontal bars leave room for longer names.
Regional revenue · long rows
One value field and one series field compare regions within each quarter.
Regional revenue · wide rows
The same numbers and chart as the long example, using one column per series.
Revenue from individual orders
Sum repeated category rows explicitly while retaining the original orders.
Net change by department
Positive and negative values share a zero baseline; these are not stacked components.
Error rate against target and average
A fixed target and a data-derived average share the value axis. Values are percentage points.
Explain a regional result
One annotation targets a whole category; another targets a specific category-and-series datum.
Signup stages
Explicit process order, rather than a data-dependent ranking.
What does the average measure?
Solid: average visible category total. Dashed: average individual value across all data. Hide Partners to see only the visible-scope statistic change.
Trend
Daily active users
Unequal date gaps occupy unequal space on a UTC scale.
Activity by platform
A null observation remains a gap; direct labels identify series and a callout explains a real datum.
Trend in a dashboard cell
An intentionally small chart with hidden axes. The full document has no title to crowd its cell.
Composition
Traffic by source
Additive traffic sources form a total over time; area stacks by default.
Revenue mix
A few positive parts of one whole, with native category and percentage labels.
Product revenue by channel
Channels are additive components of product revenue. Totals are meaningful.
Subscription mix by month
Normalize raw revenue into shares; a stable color domain preserves plan identity.
Traffic-source share over time
The same visit counts as stacked area, normalized to a whole at each date.
Issue status
Pie has its own count mode. Inner radius turns the same part-to-whole chart into a donut.
Explain one revenue segment
A pie datum anchor uses its category, while ordinary values remain native labels.
Relationship
Petal measurements by group
Independent numeric axes and categorical color; no repeated point labels.
Model size and training compute
Positive observations spanning orders of magnitude; both axes explicitly request log scales.
Store performance
Position, bubble size, and sequential color represent distinct variables with named units.
Branch performance against target
An explicit midpoint of zero distinguishes below-target from above-target performance.
Retention and acquisition payback
A coordinate annotation stays at a chosen x/y location; fixed references divide both axes.
Distribution
Tickets by category
Count individual rows; no artificial numeric value field is needed.
Range
Project phases
Horizontal start/end ranges on a date-valued axis; explicit order preserves the sequence.
Daily low and high temperature
Numeric start/end roles draw floating columns rather than scalar bars from zero.
Estimated effect by dose
Observed estimates with lower/upper fields, plus a stable-ID datum annotation.