Create a data model in Analytics
Closed beta: This feature is in closed beta and is not generally available.
Data models are available through the Early Access Program for sites on Premium and Enterprise plans. Your site must be enrolled before data models appear in Analytics. To join, submit a request at the Early Access Program sign-up form.
A data model brings together the data you want to report on. You select the tables to pull from, select the columns you need, and join them so related records line up.
You build a model once and reuse it. A single model can support many custom metrics, and each metric can appear in more than one chart.
Not yet supported in Early Access: Customers with advanced compliance requirements such as FedRAMP, HIPAA, and BYOK encryption.
Before you begin
Your site must be enrolled in the Early Access Program.
Your site must be on a Premium or Enterprise plan.
Decide which work items or assets you want to report on, and which site holds them.
To open Data Models, select App switcher, then Atlassian Analytics, then Data Models in the sidebar.
Select your site
A data model belongs to one site and can only use data from that site. Select the site at the top of the page before you start building. Only the work items and assets from the site you select are available to add to your model.
Add your data sources
Data sources are the tables your model pulls from. You can add:
the work items table, for request and work item data
object type tables from an Assets schema, for the assets your work items relate to
the Jira custom fields table, for custom field values that aren't already available as columns
Select the tables you need, then save your selection.
Select your columns
Select the columns you want in your model, then run the report to see your data.
Include every column you might want to group by or filter on later. Charts can only use columns that exist in the model, so it's quicker to add them now than to edit the model afterwards.
Spaces appear as Project Name in the column list.
Add Jira custom fields
If a custom field isn't in the column list, add the Jira custom fields table as a data source and join it to your work items table.
You can use text, number, date, and date and time custom fields. The customer-facing labels are Text Field, Short Text, Number, Date Picker, and Date Time Picker. Single select, radio button, and other complex field types aren't available yet. More field types are being added during the Early Access Program.
Join your tables
Joining tables tells Analytics how records in one table relate to records in another. Each join matches a column in one table to a column in the other.
Tables you're joining | Columns to match |
|---|---|
Work items and Jira custom fields | Join on Work item key. |
Work items and an Assets object type | Join Linked asset object key to the object type's key column. |
You can also join object types from different Assets schemas on a reference attribute. For example, you can join employees to laptops using an employee's Assigned Laptop reference field.
For more on how joins work and which type to use, see Join types.
After you join your tables, add the columns you need from the table you joined.
Add a formula column
Formula columns calculate a value from columns already in your model. You can add, subtract, multiply, divide, round a number, format a date, or find the difference between two dates.
For example, to see how long changes take to implement, add a column that calculates the difference between the planned start date and the actual completion date. You can express that difference in seconds, minutes, hours, days, weeks, months, or years, and rename the column so it's clear on a chart.
Filter your model
Filters you apply to a model also apply to every metric and chart built on it. Use a model filter when you want to narrow the data for everything downstream, such as limiting a model to a single request type.
Name and publish your model
Give your model a name that describes the data it holds, then publish it. You can rename a model later.
You must publish a model before you can build custom metrics on it.
Control who can see your model
When you share a model, you can make it visible to everyone or only selected people, and separately control who can edit it.
If someone can't access a model, they won't see the model, any metric built on it, or any chart built from those metrics.
Access to a model doesn't override existing permissions. People only see the data they already have permission to see, so two people can open the same chart and get different results.
Limitations
During the Early Access Program:
A model can only use data from one site.
Single select, radio button, and other complex custom field types aren't available as columns.
You can create up to 1,000 models.
A model can include up to 15 columns, 5 joins, and 5 data sources.
There is no row limit. The number of rows you show in a model doesn't limit the rows used in calculations.
Was this helpful?