Build your first dashboard and widget

Having data isn't enough — if it isn't visible, it rarely makes it into decisions. The fastest way for an analyst to create value is to put a small widget on top of their own dataset. This lesson takes you through that one widget end-to-end.

Create a dashboard

1. Click Dashboard in the left sidebar and choose Create dashboard. The editor (builder) opens immediately.
2. Click the title at the top and enter a name. Using your analysis topic as the name makes it easier to find later (e.g. 2026Q3-retention-summary).

At this point the dashboard is an empty canvas. Until a widget shows up, all you see is a gray plus placeholder.

Add one widget

The widget library on the left has items like:

• Visualization — Chart widgets such as bar, line, donut, and map
• Table — Shows a dataset as a plain table
• KPI — A single large number that highlights a key metric
• Markdown / image — Explanatory text

For now, drag Visualization → Bar chart onto the canvas.

Once the widget lands on the canvas, the configuration panel opens on the right.

1. Data source: Pick the dataset you created in the previous lesson.
2. Data mode: Leave it on Simple mode for now. Query mode comes in the next lesson.
3. X-axis field: Choose one categorical column (e.g. category, region, date).
4. Value field: Pick one numeric column. The default aggregate is sum or average.

Click Apply changes and the chart renders immediately. Save in the upper-right writes the dashboard, and a dashboard resource appears in the left collection tree.

Polish the widget

Selecting the same widget again surfaces additional options beyond the configuration panel:

• Title and subtitle — The label at the top of the widget. We recommend the metric (period) form (e.g. Monthly revenue (2026 Q2)).
• Sort / Top N — The two options analysts reach for most when there is a lot of data.
• Color — For category emphasis or brand color.

Touching just these three already makes a widget look like a presentable result.

Self-check

• Does the dashboard save and appear as a resource in the collection tree?
• Does the widget render meaningful data (not an empty chart)?
• Have you deliberately adjusted at least one of title, sort order, or color?

Next lesson

You did this one in simple mode, but the place analysts most often get stuck is the data configuration. The next lesson covers per-chart-type data shapes and query mode.