Group By and Split
Group by and Split by are powerful tools for gaining deeper insights. Group by changes the buckets that data is aggregated to while Split by can separate data by a secondary dimension selected.
Both of these options can be accessed via the + Group by and Split by button.
This will display the "Group by and Split" modal.
Group By is always active and the default selection is Time (Day) resulting in data being aggregated into buckets of UTC days. This is shown as a line chart over a period of time.
Group by can be changed to the following
|Time||Day, Week, Month, All|
|Global dimension||Specify a Global filter type|
|Event dimension||Specify a Event filter type. |
⚠️ Only available if an event-type with supported filters has been selected.
Group by Time
Grouping by time allow different time granularities like Day, Week and Month. It also allows a option named All which will aggregate to a single number. This can be used to show total sum of revenue over a period for example.
Group by Global Dimension
Grouping by a global dimension will aggregate data into buckets for each value found in the specified dimension type. For example grouping by Country code will aggregate into each county code found in the data and display a barchart. If the selected aggregation is Unique Users then the reported number for each country would be unique users from that country over the period selected.
Split by is an optional setting to separate the data aggregation defined by a Group by into a defined secondary (different) dimension. Only Global dimensions or Event dimensions are possible to split by.
- Splitting data grouped by time will result in multiple line charts
- Splitting data grouped by a dimension will result in a stacked barchart
Below is an example of a query with a Split by Country Code (Grouped by Time):
Group by and Split by is often combined with filtering to show only values needed. For example combining a filter for Country code equal to [UK/US] with a Group by equal to Country Code will show a bar chart for only those 2 countries.
The Limit option controls the maximum amount of unique string values returned from the data processing. If a specific dimension has more unique values in the data compared to the limit, then the top values are selected and some will be excluded.
The available options for the limit amount depends on whether a Split by active.
|Group by only||Group by can support up to 1000 unique values returned.|
|Group by and Split By||Group by supports up to 50 unique values Split by supports up to 50 values.|
This behavior is in place due to performance considerations. Enabling both Group by and Split by will cause our backend system to first query the Group by values and then for each of these query for the split values.