Charts worth stealing: The design calls BI should ship
How RJ Andrews and the dbt Charts team designed the defaults behind dbt Labs' new open source charting language, from color and typography to accessibility and AI-generated chart forms.
Overview
Teaches design best practices for bar charts and line graphs: clear comparisons, effective color and labels, and optimized design elements, demonstrated live from real YAML source code with an emphasis on dbt Charts.
What you'll take away
- 01
Most named chart types, bar, area, line, scatterplot, are variations on the same underlying grammar; dbt Charts can morph between them by changing only a couple of lines of YAML.
- 02
Default design choices are deliberate: vertical series labels sit on the right since the most recent data matters most, and every color palette is checked against roughly half a dozen types of color vision deficiency.
- 03
When a chart's data is too dense to render legibly, dbt Charts surfaces a warning during review rather than letting a broken chart reach production.
Session speaker
RJ Andrews
Fivetran