Vepr is all-in-one toolkit for building exploratory data visualization experience. Vepr implements protocol that allows declaring complex data visualizations in a scheme with small footprint and being able to transfer the visualization to different environment (e.g. server → browser, web worker → main thread).
Vepr can be installed using NPM
npm install veprAnd used as a node module
import { render, blueprint, dot } from "vepr";Or accessed via CDNs like unpkg.com right in the browser environment
<script type="module">
import { render, blueprint, dot } from "https://unpkg.com/vepr@0.0.0";
</script>Out of the box Vepr includes typing declarations for TypeScript (or TS Language Server, thus can be leveraged in JS-only environment as well).
Vepr is a research project that explores multidisciplinary nature of visual data exploration. We explore how the tech can be put to the context to gain maximum performance and fluid interactivity. Vepr is intended to be dependency-free to avoid accidental abstraction complexity. Derived code is annotated with sources wherever possible.
As a research project, Vepr is not intended to fully replace existing and established solutions. Vepr is inspired by findings and experience of many developers involved in creating these tools.
Vega introduces declarative format for interactive visualizations. Being a plain JSON schema, the visualization declaration can be generated in variety of environments and then visualized in the browser. Many different data viz libraries are built on top of Vega grammars and are used a lot in Python notebooks. In comparison, Vepr implements a format that does not expose underlying data so all data-related computations (aggregations, filtering, sorting) can be done outside of browser environment (for performance purpose) thus significantly reducing the declaration size while keeping the resulting visualization interactive and adaptive.
Plot is an exploratory visualization library optimized for incremental progress. It is very ergonomic and has great defaults that provide faster feedback cycle. Resulting visualization can be also well describe by the code generating it, which is used efficiently in notebook environment. In comparison, Vepr is created with software engineers in mind that don't necessarily know the nature of the data they work with (e.g. when creating a universal dashboard for customers). Vepr may not be as expressive or may require explicit definitions for things that would otherwise be inferred.
While Vepr is in active development phase, please consider using stable and well known solutions.