dbt Labs has released dbt Charts, an open source charting language, and detailed its design choices at a dbt Summit 2026 breakout session. Chartographer RJ Andrews argued that named chart types like bars, areas, lines and scatterplots are variations on one underlying grammar, morphing a default bar chart into each form with a couple of lines of YAML.

The default bar chart places series labels on the right, where the newest data sits, and switches between monthly and quarterly axis labels based on data density, showing the year only at a series start or change. Three themes ship: Clarity, Neon dark mode and Paper, a cream background inspired by the Financial Times and the Economist.

Andrews said most BI tools pair blue with high-contrast orange or red, drawing attention to the contrast rather than the data, so dbt Charts keeps palettes harmonious for charts with few series, holds category colors across a dashboard and uses darker shades for single-series charts. Palettes were checked against about half a dozen types of color vision deficiency, weighted by population prevalence. Pattern fills are built but not yet public.

Dense data triggers a warning during review instead of a broken chart reaching production. Tables align decimals without adding false precision, and dashboard filters update the URL so views can be shared. An agent can generate unconventional forms like lollipop or bump charts from plain-language requests.

Source: dbt Labs Blog