visgl/deck.gl
Developer ToolsA GPU-powered WebGL2/WebGPU framework for building high-performance large-scale geospatial and data visualizations.
What it does
deck.gl is a visualization framework that renders large datasets on the GPU using WebGL2/WebGPU. It maps input data — typically an array of JSON objects — onto a stack of visual layers (icons, polygons, text, and more) and displays them through views such as a map, first-person camera, or orthographic projection. It bundles performant rendering and updates, interactive picking/highlighting/filtering, cartographic projections with major basemap integrations, and a tested layer catalog. Layers expose flexible APIs and the core classes are extendable for custom needs.
It targets developers building interactive, data-heavy map or spatial visualizations on the web, and offers flavors for script-tag usage, npm with plain JS or React, and Python via pydeck.
Quick start
With npm and React:
npm install deck.gl
Then follow the React getting-started guide and the React examples. For plain JS, use the standalone guide and pure-JS examples.
A minimal script-tag route also exists: load the bundle, then follow the scripting API guide.
<script src="https://unpkg.com/deck.gl@latest/dist.min.js"></script>
The README links to full script-tag examples. The Python route installs pydeck and has its own installation guide and layer examples.
Constraints
Rendering assumes a WebGL2/WebGPU-capable browser environment; map-style views depend on integration with a basemap provider. The quoted install commands only cover package installation — select a view, layer set, and basemap suitable for the target deployment. The project is MIT-licensed and is part of vis.gl under the OpenJS Foundation; contribution rules are in CONTRIBUTING.md.
Explore
The API documentation, website demos with source links, interactive playground, and CodePen demos are the fastest paths to seeing layer behavior before wiring in your own data.
HotRepo's take
On the 2026-11-19 daily TypeScript list, deck.gl recorded 4 new stars, a best position of 12, and totals of 14,774 stars and 2,306 forks. That is a mature project's measured position rather than a growth story; the strongest signals here are its layer/view architecture, multiple language bindings, and long-running backing by major map and data companies.
Written by AI from the project's README, updated . The project's own docs are the source of truth.