Vega Visualization Grammar

Vega is a free and open-source visualization grammar, a declarative format for creating, saving, and sharing interactive visualization designs. Created by the University of Washington Interactive Data Lab (UW IDL, led by Jeffrey Heer) and first released in 2013, Vega is the foundation for several popular visualization tools including Vega-Lite and the Lyra visualization design environment. Key features: declarative grammar: visualizations are described as JSON specifications rather than imperative code. A Vega specification defines data sources, scales, axes, legends, marks (visual elements like rectangles, circles, paths), signals (dynamic variables), and interaction event handlers. This declarative approach enables reproducible, serializable, and shareable visualizations. Vega-Lite: a higher-level grammar built on top of Vega, providing a more concise syntax for common chart types (bar, line, area, scatter, histogram, boxplot). Vega-Lite automatically generates Vega specifications, handling scale domain/range, axis configuration, and layout. Used by many tools including Altair (Python), vl-plot (R), and Observable. Data flow: Vega's data flow graph model processes data through transformations (filter, aggregate, bin, sort, calculate) and pipes results to marks for rendering. The reactive dataflow runtime efficiently updates visualizations when data or parameters change. Interactivity: built-in support for interactive features including tooltips, zooming, panning, brushing, linking, and parameter widgets. Selections enable interactive filtering and highlighting. Rendering: SVG and Canvas rendering backends. The SVG backend produces crisp, scalable, and exportable graphics. Canvas backend provides higher performance for large datasets. Multiple language support: specifications can be authored in JSON, Python (via Altair), R (via vl2), JavaScript, and compiled to static images. Cross-platform JavaScript/TypeScript codebase. BSD-3-Clause.