Power BI + Graphic Walker Online: A Tableau-like Interface for Vega-Lite

Published Categorized as Custom Visualisations, DataViz, Deneb No Comments on Power BI + Graphic Walker Online: A Tableau-like Interface for Vega-Lite

If you’ve ever wanted the flexibility of Deneb but felt overwhelmed by its code-first approach, Graphic Walker Online could be a good bridge. Graphic Walker is posited as Tableau-style front end for Vega-Lite — ideal for beginners or anyone who wants to visually explore data and then export Vega-Lite specs for use elsewhere, including Power BI.

Graphic Walker Online makes it easy to get started, even if you don’t have a dataset handy.

Visit Graphic Walker Online https://graphic-walker.kanaries.net/

And begin by selecting a dataset.

Create Dataset > select a sample dataset (such as Titanic passengers or Iris flowers) or upload your own.

The following screen allows you to review and filter data as needed, then load

The visual canvas will look familiar to Tableau users. It is easy to drag and drop dimensions and measures to the “X” and “Y” fields — just like Tableau.

It is also easy to add filters, and color, shape and size encodings as needed.

Different chart types like bar, line, scatter, or area can also be explored with one click.

Re-use in Power BI

Once your chart looks right:

Select “Export Code” button on the ribbon to open the underlying Vega-Lite specification.

Copy the JSON code between the outer braces.

In Power BI, install the Deneb custom visual from AppSource.

Select new template, with or without Power BI theming.

Within the specification editor, overtype the ‘ “layer” : [ ] ‘ line with the copied the Vega-Lite spec.

Bind Power BI data fields to match your Vega-Lite spec (you may need to adjust the field names).

Ensure your dataset has the same structure as what you built in Graphic Walker.

Keep an eye on field types: Vega-Lite is strict about data types (e.g., quantitative vs. nominal).

Considerations:

One of the limitations of Graphic Walker is that it appears to have a limited Vega-lite featureset (e.g. ability to layer or use transformations like regression).

However,

This is a great option to quickly explore your data with a Vega-lite GUI, then save the resultant output, which can be developed further.

  • No coding required: You can explore data and create visualizations without writing a single line of Vega-Lite code.
  • Immediate feedback: Like Tableau, every drag-and-drop updates the chart instantly.
  • Learning by doing: Once you export the spec, you can study how your actions translated into Vega-Lite syntax.

Happy Vizin’

Kez

Leave a comment

Your email address will not be published. Required fields are marked *