selecting-vizro-charts

Match the chart to your data question: bar charts for comparing categories, line charts for trends over 12+ time points, pie charts only for 2-5 part-to-whole slices, histograms for distributions, and scatter plots for correlations. Avoid 3D charts, dual Y-axes, and pies with 6+ slices.

Plotly Express does not aggregate data, so passing detail-level rows stacks individual rows as separate rectangles instead of summing values. Write a @capture("graph") custom chart function that aggregates the data inside before plotting.

No. Plotly charts and KPI cards automatically use Vizro template color schemes, so do not set marker_color, hex codes, or color_discrete_sequence unless the user explicitly requests it. Only AG Grid needs manual colors via vizro.themes palettes.

Use vm.AgGrid with figure=dash_ag_grid(data_frame=...) from vizro.tables. Never use vm.Table, Dash DataTable, or fake tables with Plotly heatmaps. For custom column logic, write a @capture("ag_grid") function with data_frame as the argument name.

Use @capture("graph") when you need aggregation or sorting before plotting, update_layout or update_traces calls, reference lines, parameter-driven logic, dual-axis charts, or multi-trace go.Figure compositions. Simple px charts expressible in YAML do not need custom functions.

No. Use the built-in kpi_card or kpi_card_reference from vizro.figures inside a Figure model, with titles in the figure args. Custom chart implementations are only acceptable when the KPI is strictly impossible with built-ins, such as dynamically displaying text.