About data blending

Join information from multiple sources to get a more unified view of your data.

By default, charts in Data Studio get their information from a single data source. Blending lets you create charts based on multiple data sources, called a blended data source. For example, you can blend two different Google Analytics data sources to track the performance of your app and website in a single visualization.

Blending can reveal valuable relationships between your data sets. Creating blended charts directly in Data Studio removes the need to manipulate your data in other applications first, saving you time and effort.

Get started quickly!

Watch a video demonstration.

Get started with data blending

See a sample report with blended data.

Click the example report below to see how you can join data from 2 different spreadsheets.

Blending demonstration report


How blending works

Blended data sources are created by joining the records in one data source to the records of up to 4 other data sources. To join the data, each data source in the blend must share one or more dimensions, known as a join key (or just a key, for short). Blended data sources include all the records from the leftmost data source in the Blend Data panel, and the matched records from the data sources to the right.

For example, the screenshot below shows the Blend Data panel with 3 data sources: Website A, Website B, and Mobile Apps A. The join key is the Source dimension. The blended data source created from this includes all the records from Website A, along with any records from Website B and Mobile Apps A that share the same Source values as Website A.

Example of joining 3 data sources.


Technically, a blended data source is the product of a left outer join operation. In a left outer join of table A and table B, the result is all the records of Data source A and those records in Data source B that share the same key values.

The diagram below illustrates a left outer join of data sources A and B. The blended data source includes all the records contained by the green circle.

Venn diagram of a left outer join.

Blended data sources only exist in reports: you won't see them in your DATA SOURCES Home page. You can, however, manage them via the Resources > Manage blended data menu.

Blend using multiple dimensions

You can blend data sources using multiple dimensions as the join key. Each data source in the blend must have the same set of dimensions used in the key. Here's an example:

Example of a multi key join.

In this blend, only the records from Store Orders that match both Sales Rep ID and Region in Sales Reps will be included in the data source.

Blend a data source with itself

You can blend a data source with itself. To do this, add the same data source more than once in the Blend Data panel.

For example, the Google Analytics connector contains metrics for 1 day active users, 7 day active users, and 28 day active users. But, due to a limitation of Analytics, you can only have one of these metrics in a chart at a time. By joining the same Analytics data source with itself, you can add each of these metrics to the blended data source. You can then compare each of these active users metrics in the same chart. Cool, huh?

Manage blended data sources

Blended data sources in a report are listed in the DATA tab of the properties panel, under Component Data Sources.

You can check the status of and remove blended data sources using the Resources > Manage blended data menu.

Removing a blended data source will break any charts that depend on it, but will not remove the underlying data sources themselves. Any data sources you've added will still be attached to the report.

Limits of blending data

Blended data sources belong to the report in which they were created. To reuse a blended data source in another report, copy and paste a component with blended data into the new report.

You can blend up to 5 data sources in a chart.

Blending is only available in reports.

Creating calculated fields in blended data sources is not currently supported.

Was this article helpful?
How can we improve it?