Chart transformations

Chart transformations are an additional transformation configuration that applies specifically to a graph chart. While the dataset’s transformations shape how the data appears in a table view, chart transformations shape how the data is prepared for the graph.

​

Why a separate layer?

A single dataset might power both a table and a graph on the same dashboard. The table might need to show individual rows, while the graph needs data grouped by month. Chart transformations let you configure the graph’s data shaping independently without affecting the table.

​

What you can configure

Chart transformations support the same operations as dataset-level transformations:

• Group by: group rows by columns to create the axis categories or time series.
• Aggregate: apply sum, average, count, and other functions to numeric columns.
• Filter: narrow the data for the graph.
• Sort: control the order of data points.
• Hide columns: exclude columns not needed by the graph.
• Row limit: cap the number of data points.

​

Example

Suppose your dataset returns daily sales by product: The dataset’s transformations could be left unmodified so the table shows all rows. The chart’s transformations could group by date and sum amount to feed a line chart showing total daily sales.

​

Related

• Transformations, the dataset-level layer.
• Graph charts, configuring graph visualizations.
• Charts overview