uvdl

Uvdl, short for Universal Visualization Description Language, is a proposed domain-specific language intended to describe data structures, transformations, and visual encodings within data visualization workflows. It is primarily discussed in academic and design contexts as a means to standardize how datasets are described and how visual representations are mapped from data.

Overview and purpose

Uvdl aims to provide a human-readable, machine-parseable format that can be used to share visualization specifications across tools and platforms. A uvdl document typically defines datasets, transformation steps, and the visual encodings applied to those data. The language emphasizes clarity and modularity to facilitate interoperability between analysis pipelines, visualization engines, and collaborative workflows.

Design and features

Uvdl uses a hierarchical syntax that supports common data types, columns, and basic operations such as filter, aggregate, and join. Visual encodings specify elements like marks, axes, color scales, and simple interactions. The design prioritizes explicit declarations of data provenance, transformation steps, and rendering targets to reduce ambiguity between stages of a visualization pipeline.

Example

dataset:

name: sample

columns:

- id: int

- value: float

transform:

- type: filter

condition: value > 0

viz:

- type: line

x: id

y: value

color: none

Status

Uvdl remains a theoretical or experimental concept without an officially ratified standard. Prototype implementations appear in educational and research tools, serving as a reference for exploring how visualization specifications might be exchanged between systems.

See also

Vega-Lite, JSON, YAML.