UseR! 2019 Presentation

A talk about the vegawidget organization

It’s very exciting to be presenting at UseR! 2019 in Toulouse, France.

This presentation [] incorporates contributions from the entire vegawidget organization:

• Ian Lyttle
• Haley Jeppson
• Alicia Schep
• Stuart Lee
• Wenyu Yang
• Heike Hofmann

This post also serves as an announcement for the CRAN release of vegawidget 0.2.1, which incorporates Vega-Lite 3.3.0.

The vegawidget package plays a role within a larger process. Here’s how we see the end-to-end process of building and deploying a Vega-Lite chart using R:

Code icons made by Smashicons from www.flaticon.com

¹

Working backwards from the end-user, we see three steps: the Vega-Lite library itself displays a chart in a browser, rendered from a JSON specification, which is composed by an R programmer.

• Vega-Lite

  • provides an interactive grammar-of-graphics
  • JavaScript library that runs in the browser

• Rendering

  • vegawidget: htmlwidget package to render Vega-Lite specifications to Vega-Lite; also to provide interactive access (including Shiny) to Vega-Lite charts.

• Composing

  These packages are created by members of the vegawidget organization:

  • altair: wraps Python package Altair, which covers entire Vega-Lite API.

  • vlbuildr: covers entire Vega-Lite API, with pipeable functions.

  • ggvega: translator from ggplot2 to Vega-Lite.

Going forward, we expect to focus more-and-more on the packages to compose specifications, while the vegawidget package renders in the “background”.