billboard.js: Naver's D3 chart library, now with a Canvas lane

The chart library named after the charts

billboard.js is a re-usable, easy interface JavaScript chart library based on D3.js, and its name is explained in one line, it comes from the famous billboard chart everybody knows, the music ranking rather than the roadside sign. The project is developed under Naver, Korea's search giant, is MIT licensed, and its repository carries a wiki page titled Why we decided to start billboard.js explaining the origin, alongside a migration guide from C3.js that tells the genealogy more plainly, this is the D3-charting-for-humans idea that C3 pioneered, restarted with fresh foundations. Version history marks the care taken with that lineage, v3 brought D3 v6 support and a candlestick chart type, and v4 added the Canvas rendering mode, each with its own changelog and migration notes.

v4's Canvas lane for high-density axes

The headline change of the current major version is an opt-in Canvas rendering mode aimed at high-density axis charts, the scatter and line plots where thousands of SVG nodes start to hurt. The project is unusually disciplined about stating the mode's boundary, the Canvas notes document which chart types are supported, how ESM import usage differs, what happens to interactions, and what the SVG parity scope is, the four questions anyone evaluating a secondary renderer actually has. The repository also carries a CANVAS_THEME_SELECTORS.md file, evidence that theming had to be rethought for the second renderer rather than assumed to carry over, and the package exports map exposes a separate ./canvas entry point with its own type definitions, so the Canvas build is a deliberate sub-package rather than a flag on the main one.

Five themes and one packaged build with a warning

Theming is a CSS decision, load one of the provided theme files instead of the default stylesheet and the charts change clothes, with five themes shipping in the distribution, datalab, dark, insight, graph and modern, each in plain and minified form. Distribution itself is generous, uncompressed and compressed builds, latest and version-pinned URLs, all served from the project's GitHub Pages, so a team can standardize on an exact version by URL without a package manager. There is also a packaged variant, billboard.pkgd.js, that bundles D3.js inside for one-script convenience, and the README attaches an honest warning to it, the packaged version is not an official distribution, existing to provide an easy way to load billboard.js with dependencies, which tells you the maintainers would rather you manage D3 yourself.

Nightly, next and semantic-release

Three release channels run in parallel. The stable channel installs with pnpm:

$ pnpm add billboard.js

It stands at 4.1.0, released 2026-09-16. The next channel is the release candidate:

$ pnpm add billboard.js@next

And the recent 4.1.0-next.2 and 4.1.0-next.3 tags show that candidate track being exercised ahead of a stable cut. The nightly channel builds from master on every push, with version strings encoding the build datetime in the form x.x.x-nightly-yyyymmddhhmmss, installable either by pointing package.json at naver/billboard.js#nightly or by cloning the git URL with the branch ref. Behind all of it sits semantic-release, the automation that derives versions from commit messages, which explains how a solo-looking maintenance effort sustains multi-channel releases, and the repository was last pushed on 2026-09-28.

The exports map as a feature list

The package manifest is worth reading as documentation, because its exports map enumerates the surfaces a modern consumer can address, the root library, the ./canvas entry with its own build, the ./react entry for the React component with dedicated types, ./dist/plugin/* for the plugin builds, and the raw ./src/Plugin/* path for importing plugin source. Module formats are layered deliberately, an ESM build produced with rolldown, a CommonJS path generated by a custom script, and the webpack-produced dist files for script-tag use, with types wired per entry through typesVersions. The React component deserves the emphasis the manifest gives it, chart libraries live or die by framework integration, and here it is a first-class sub-package rather than a community afterthought.

A documentation set that includes a benchmark

The supporting material goes beyond the usual API reference and demo. The demo site at naver.github.io/billboard.js covers the supported chart types with live examples and source, the API documentation is generated from the code with jsdoc, and there is an actual benchmark page comparing rendering performance, plus a benchmark directory in the repository. The wiki layers on the judgment material, a comparison table with other libraries, a who's-using page, third-party applications, and technical articles on bundling for legacy browsers and generating chart images in a Node.js environment, the server-side rendering question every chart library eventually faces. Questions are routed to Stack Overflow under the billboard.js tag or to issues labeled question, a split that keeps the bug tracker focused.

TypeScript source and two test rigs

Under the hood the library is written in TypeScript, formatted with dprint, linted with its own eslint configuration, and tested with both vitest and mocha configured side by side, a transition state several long-lived projects pass through when modernizing their test stack. The build pipeline spans webpack for the dist artifacts and rolldown for ESM, with separate production, packaged, theme, plugin and type-generation steps, and a Storybook instance with committed stories serves as living documentation of the visual surface. Engineering notes persist as first-class files, PERFORMANCE.md, MODULE_IMPORTS.md and CANVAS_THEME_SELECTORS.md, and AUTHORS.txt records contributors, the posture throughout being that of a library maintained by people who expect to keep maintaining it.

Editorial conclusion

Choose billboard.js when you want D3-grade charting through a friendly configuration API, with the option of Canvas rendering for dense axis data and CSS-swappable themes out of the box. Choose C3.js if you are already on it, though the project publishes a migration guide precisely because billboard.js began as that idea done fresh, or Chart.js when its all-in-one bundle matters more than D3 lineage. Verify first whether your chart types fall inside the documented Canvas parity scope before relying on that mode, prefer the official dist files over the packaged D3-bundled build the README flags as unofficial, and consult the comparison wiki page for the feature-by-feature table before committing.

Frequently asked questions

Official sources

1. License: MIT
2. naver/billboard.js on GitHub
3. Project website
4. README
5. Releases