A charting library for Typst built entirely with native primitives (rect, circle, line, polygon, place). No external dependencies required.

Gallery

All 50+ chart types across 8 pages — see examples/showcase.typ for the source:

Examples

Shared datasets in data/ used by both demo and showcase:

• data/sales.json — SaaS startup metrics
• data/codebase.json — Open source project stats
• data/league.json — Soccer league data
• data/rpg.json — D&D campaign tracker
• data/words.json — Data visualization vocabulary

Compile any example directly with typst compile examples/demos/demo-bar.typ.

Features

• 50+ chart types for data visualization
• JSON data input — load data directly from JSON files
• Theme system — golden-ratio proportional scaling from two seeds (base-size, base-gap), preset themes, custom overrides, and with-theme() for document-wide defaults
• Nice ticks — D3-style axis tick algorithm produces round, evenly-spaced values; auto-detects decimal precision from step size
• Smart label placement — automatic fit detection, font shrinking, and greedy deconfliction for overlapping labels
• Layout primitives — shared utilities for label density, font scaling, and label placement
• Annotations — overlay reference lines, bands, labels, content, points, error bars, and rectangles on any Cartesian chart
• Error bars — native errors: parameter on bar and line charts (symmetric or asymmetric)
• Outliers — native outliers: field on box-plot boxes
• Relative widths — use width: 100% for responsive charts inside containers and grids
• Dashboard primitives — card, compact-table, alert, badge, separator, and dashboard-layout for report layouts
• Customizable — colors, sizes, labels, legends
• Pure Typst — no packages or external tools needed

Chart Types

Bar Charts

• bar-chart - Vertical bar chart
• horizontal-bar-chart - Horizontal bar chart
• grouped-bar-chart - Side-by-side grouped bars
• stacked-bar-chart - Stacked bar segments
• grouped-stacked-bar-chart - Groups of stacked segments side by side
• lollipop-chart - Vertical stem + dot (cleaner bar alternative)
• horizontal-lollipop-chart - Horizontal stem + dot
• diverging-bar-chart - Left/right bars from center axis

Line & Area Charts

• line-chart - Single line with points
• multi-line-chart - Multiple series comparison
• dual-axis-chart - Two independent Y-axes
• area-chart - Filled area under line
• stacked-area-chart - Stacked area series

Circular Charts

• pie-chart - Pie chart with legend
• pie-chart (donut mode) - Donut/ring chart
• radar-chart - Spider/radar chart

Scatter & Bubble

• scatter-plot - X/Y point plotting
• multi-scatter-plot - Multi-series scatter
• bubble-chart - Scatter with size dimension, smart label deconfliction (inside/outside with leaders)
• multi-bubble-chart - Multi-series bubble chart

Gauges & Progress

• gauge-chart - Semi-circular dial gauge
• progress-bar - Horizontal progress bar
• circular-progress - Ring progress indicator
• ring-progress - Concentric fitness rings (Apple Watch style)
• progress-bars - Multiple comparison bars

Sparklines (inline)

• sparkline - Tiny line chart for tables and text
• sparkbar - Tiny bar chart
• sparkdot - Tiny dot chart

Heatmaps

• heatmap - Grid heatmap with color scale
• calendar-heatmap - GitHub-style activity grid
• correlation-matrix - Symmetric correlation display

Statistical

• histogram - Auto-binned frequency distribution
• waterfall-chart - Bridge/waterfall chart with pos/neg/total segments
• funnel-chart - Conversion funnel with percentages, auto external labels for narrow sections
• box-plot - Box-and-whisker distribution plot
• treemap - Nested rectangles for hierarchical data
• slope-chart - Two-period comparison with connecting lines
• bullet-chart - Compact gauge with qualitative ranges and target
• bullet-charts - Multiple bullet charts stacked vertically

Proportional & Hierarchical

• waffle-chart - 10×10 grid of colored squares for proportions
• sunburst-chart - Multi-level hierarchical pie with nested rings
• parliament-chart - Semicircle dot layout for seat visualization

Comparison & Ranking

• bump-chart - Multi-period ranking chart
• dumbbell-chart - Before/after dot comparisons with connecting lines
• radial-bar-chart - Circular bars radiating from center

Distribution

• violin-plot - Kernel density estimation with mirrored polygon

Flow & Timeline

• sankey-chart - Flow diagram with curved bands between nodes
• gantt-chart - Timeline bar chart for project scheduling
• timeline-chart - Vertical event timeline with alternating layout
• chord-diagram - Circular flow diagram with chord bands

Dashboard

• metric-card - KPI tile with value, delta, and sparkline
• metric-row - Horizontal row of metric cards
• card - Themed container with optional title and description
• compact-table - Dense data table with header styling and highlight column
• alert - Info/warning/error/success notification block with left border accent
• badge - Inline colored pill (default/secondary/destructive/outline/success)
• separator - Themed horizontal rule
• dashboard-layout - Grid layout helper for multi-row dashboard pages
• word-cloud - Weighted text layout sized by importance

Annotations

Overlay extra elements on any cartesian chart (bar, line, area, scatter, bubble, box-plot, histogram, violin, waterfall, dual-axis, …) via the annotations: parameter. All coordinates are in data space.

• h-line — horizontal reference line (target, average, threshold)
• v-line — vertical reference line
• h-band — horizontal shaded region (goal zone, range)
• label — text label at a data point
• content — place arbitrary Typst content at (x, y); anchor: controls positioning (top-left, top, center, bottom-right, etc.)
• point — circle marker at (x, y) with configurable radius, fill, stroke
• errorbar — vertical (default) or horizontal (orientation: "h") error bar with caps
• rect — rectangle in data coordinates with fill, stroke, opacity

#bar-chart(data, annotations: (
  (type: "h-line", value: 160, dash: "dashed", label: "target"),
  (type: "content", x: 2, y: 165, body: [★ best], anchor: "bottom"),
  (type: "errorbar", x: 0, low: 115, high: 125),
))

Error bars (native)

Bar and line charts accept an errors: parameter. Axis range auto-extends to cover the error extremes.

#bar-chart(data, errors: (10, 12, 8, 15))                       // symmetric
#line-chart(data, errors: ((0.5, 0.7), (0.3, 0.4), (1, 2)))     // asymmetric (low, high)

Box-plot outliers

Each box can carry an optional outliers: array. Outlier values are drawn as small dots and extend the y-axis range.

#box-plot((
  labels: ("A", "B"),
  boxes: (
    (min: 20, q1: 35, median: 50, q3: 65, max: 80, outliers: (95, 5)),
    (min: 15, q1: 30, median: 45, q3: 60, max: 75),
  ),
))

Installation

#import "@preview/primaviz:0.6.0": *

Usage

#import "@preview/primaviz:0.6.0": *

// Load data from JSON
#let data = json("mydata.json")

// Create a bar chart
#bar-chart(
  (
    labels: ("A", "B", "C", "D"),
    values: (25, 40, 30, 45),
  ),
  width: 300pt,
  height: 200pt,
  title: "My Chart",
)

// Create a pie chart
#pie-chart(
  (
    labels: ("Red", "Blue", "Green"),
    values: (30, 45, 25),
  ),
  size: 150pt,
  donut: true,
)

// Create a radar chart
#radar-chart(
  (
    labels: ("STR", "DEX", "CON", "INT", "WIS", "CHA"),
    series: (
      (name: "Fighter", values: (18, 12, 16, 10, 13, 8)),
      (name: "Wizard", values: (8, 14, 12, 18, 15, 11)),
    ),
  ),
  size: 200pt,
  title: "Character Comparison",
)

Theming

Every chart function accepts an optional theme parameter. Themes control colors, font sizes, grid lines, backgrounds, and other visual properties.

Using a preset theme

#import "@preview/primaviz:0.6.0": *

#bar-chart(data, theme: themes.dark)

Document-wide theme

Use with-theme() to set a default theme for all charts in a block — no need to pass theme: to every chart. Explicit per-chart theme: parameters still override.

// Block wrapper
#with-theme(themes.dark)[
  #bar-chart(data)       // uses dark theme
  #pie-chart(data2)      // uses dark theme
  #line-chart(data3, theme: themes.minimal) // explicit override wins
]

// Or as a show rule for the entire document
#show: with-theme.with(themes.dark)

Scaling with seeds

All font sizes and spacing are derived from two seed values via golden-ratio (φ = 1.618) powers. Change the seeds to scale everything proportionally:

#bar-chart(data, theme: (base-size: 10pt, base-gap: 8pt))  // larger text and spacing
#bar-chart(data, theme: (base-size: 5pt, base-gap: 3pt))   // compact

Custom overrides

Pass a dictionary with only the keys you want to change. Unspecified keys fall back to the active theme (global or default). Partial overrides merge onto the global theme set by with-theme(), so theme: (show-grid: true) inside a with-theme(themes.dark) block gives you dark + grid:

#bar-chart(data, theme: (show-grid: true, palette: (red, blue, green)))
#bar-chart(data, theme: (tick-digits: 2))  // force 2 decimal places on axis ticks

Theme from JSON

Build a theme from a JSON tokens file (e.g., exported from a CSS design system):

#let tokens = json("tokens.json")
#let my-theme = theme-from-json(tokens.light)
#let my-dark = theme-from-json(tokens.dark)

#show: with-theme.with(my-theme)

Expected JSON format: palette (array of hex strings), text-color, text-color-light, text-color-inverse, background (hex or null), border-color, border-radius (number in pt).

Extracting themes from CSS

Two optional helper scripts — a Python version (scripts/extract-theme.py, uv + coloraide) and a TypeScript version (scripts/extract-theme.ts, Bun + culori) — live in the project repository. They convert CSS design tokens (--chart-1…N, --foreground, --background, etc.) into primaviz theme files in .typ and/or .json format, handling oklch/hsl/rgb/hex color spaces with alpha blending.

⚠️ Security note: These scripts are not part of the published package archive. Before running them on your machine, please read the source in the links above to understand what they do — don’t blindly execute third-party scripts.

Both accept the same flags: --out-dir, --format (typst/json/both), --name, --dark-selector. See the repo’s README for the full run-through.

Custom theme keys

Themes support passthrough of custom keys not in the default theme. This lets you extend the theme system for your own components:

#let my-theme = (
  palette: (red, blue, green),
  card-fill: rgb("#f5f5f5"),  // custom key — preserved and accessible
)
#show: with-theme.with(my-theme)

Available presets

Data Formats

Simple data (labels + values)

(
  labels: ("Jan", "Feb", "Mar"),
  values: (100, 150, 120),
)

Multi-series data

(
  labels: ("Q1", "Q2", "Q3"),
  series: (
    (name: "Product A", values: (100, 120, 140)),
    (name: "Product B", values: (80, 90, 110)),
  ),
)

Scatter/bubble data

(
  x: (1, 2, 3, 4, 5),
  y: (10, 25, 15, 30, 20),
  size: (5, 10, 8, 15, 12),  // for bubble chart
)

Heatmap data

(
  rows: ("Row1", "Row2", "Row3"),
  cols: ("Col1", "Col2", "Col3"),
  values: (
    (1, 2, 3),
    (4, 5, 6),
    (7, 8, 9),
  ),
)

Color Palette

The default theme uses Tableau 10 colors. You can access colors from any theme via the get-color function:

#import "@preview/primaviz:0.6.0": get-color, themes

// Default palette
#get-color(themes.default, 0)  // blue
#get-color(themes.default, 1)  // orange
#get-color(themes.default, 2)  // red

// Or use a theme preset
#get-color(themes.dark, 0)  // cyan

Repository & contributing

Source, issue tracker, development scripts, and project structure live at github.com/phiat/primaviz. Bug reports and pull requests are welcome there.

License

MIT