All products

Create charts from SQL data with AI

Express a visualisation alongside the data query using ggsql. Choose layers, scales, facets and labels that fit the underlying data and the comparison you need.

To create charts from SQL data with AI, give the ggsql skill the table structure and the question behind the report. If overall orders have fallen, a weekly series for each branch can reveal whether the drop is widespread. The query should state which dates and orders count before colour or layout enters the discussion.

The agent writes the visualisation alongside the data selection. Geometries describe what to draw, scales control how values appear and facets can separate categories that would otherwise overlap. Keeping those choices in the query makes a refreshed chart easier to reproduce. Run it with a compatible version of the tool and compare a few values with the source records.

What it does

  • Select the data needed for a chart.
  • Build line, point or bar layers.
  • Separate groups into comparable panels.
  • Set axis labels, colour scales and presentation details.
  • Check aggregation before drawing conclusions.

Install and use

  1. Install the skill in a compatible agent. To render, use a compatible ggsql runtime and authorised access to the relevant dataset.
  2. Ask “Show weekly orders by branch from this table. Use separate panels if the time series overlap and label the counting unit clearly.”
  3. Run the query and verify one week against the underlying totals.

FAQ

Source

Made by George Stagg. Original skill: ggsql, published on GitHub under the MIT licence.

View on GitHub