Customer Journey Analytics BI extension

The Customer Journey Analytics BI extension enables SQL access to the data views that you have defined in Customer Journey Analytics. Your data engineers and analysts might be more familiar with Power BI, Tableau Desktop, or other business intelligence and visualization tools (further referred to as BI tools). They can now create reporting and dashboards based on the same data views that Customer Journey Analytics users are using when creating their Analysis Workspace projects.

Adobe Experience Platform Query Service is the SQL interface to data available in the data lake of Experience Platform. With the Customer Journey Analytics BI extension enabled, the functionality of Query Service is extended to see your Customer Journey Analytics data views as tables or views in a Query Service session. As a result, business intelligence tools that use Query Service as their PostgresSQL interface benefit seamlessly from this extended functionality.

The main benefits are:

Prerequisites

To use this functionality, you can use expiring or non-expiring credentials to connect BI tools to the Customer Journey Analytics BI extension. The Credentials guide provides more information on setting expiring credentials or non-expiring credentials.
Below are additional steps to set up CJA Permissions

Expiring credentials

To use expiring credentials, you can:

Or you can:

Non-Expiring credentials

To use non-expiring credentials:

See Customer Journey Access Control for more information, specifically the Product Admin additional permissions and Customer Journey Analytics Permissions in the Admin Console.

Usage

To use the Customer Journey Analytics BI extension functionality, you can either use SQL directly or use the drag and drop experience available in the specific BI tool.

SQL

You can use the functionality directly in SQL statements using either the Query Editor or a standard PostgresSQL command-line interface (CLI) client.

See the Query Editor UI guide for more information.

BI tools

Currently, the Customer Journey Analytics BI extension is supported and tested for Power BI and Tableau Desktop only. Other BI tools using the PSQL interface might work as well, but are not yet supported officially.

See Connect clients to Query Service for an overview of and more information on the various tools available.

See Use cases on how to accomplish a number of use cases using the Customer Journey Analytics BI extension

Functionality

By default, your data views have a table-safe name generated from their friendly name. For example, the data view named My Web Data View has the view name my_web_data_view. You can define a preferred name to use in your BI tool for your data view. See Data view settings for more information.

If you want to use the data view IDs as the table names, you can add the optional CJA_USE_IDS setting to your database name when connecting. For example, prod:cja?CJA_USE_IDS shows your data views with names like dv_ABC123.

Data governance

The data governance-related settings in Customer Journey Analytics are inherited from Adobe Experience Platform. The integration between Customer Journey Analytics and Adobe Experience Platform Data Governance allows for labeling of sensitive Customer Journey Analytics data and enforcement of privacy policies.

Privacy labels and policies that were created on datasets consumed by Experience Platform can be surfaced in the Customer Journey Analytics data views workflow. Therefore, data queried using the Customer Journey Analytics BI extension show appropriate warnings or errors when not complying with the privacy labels and policies defined.

List data views

In the standard PostgreSQL CLI, you can list your views using \dv

prod:all=> \dv
                       List of relations
 Schema |                    Name                    | Type |  Owner
--------+--------------------------------------------+------+----------
 public | my_web_data_view                           | view | postgres
 public | my_mobile_data_view                        | view | postgres

Nested versus flattened

By default, the schema of your data views uses nested structures, just like the original XDM schemas. The integration also supports the FLATTEN option. You can use this option to force the schema for the data views (and any other table in the session) to be flattened. Flattening allows for easier use in BI tools that don’t support structured schemas. See Working with nested data structures in Query Service for more information.

Defaults and limitations

The following additional defaults and limitations apply when using the BI Extenion:

Supported SQL

See Query Service SQL reference for the full reference on what type of SQL is supported.

See the table below for examples of the SQL you can use.

SELECT * FROM dv1 WHERE 1=0

SELECT dim1, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY dim1

SELECT dim1, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’ AND
 filterId = ‘12345’
GROUP BY dim1

SELECT dim1, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’ AND
 AND (dim2 = ‘A’ OR dim3 IN (‘X’, ‘Y’, ‘Z’))
GROUP BY dim1

SELECT dim1, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY dim1
HAVING m1 > 100

SELECT DISTINCT dim1 FROM dv1

SELECT dim1 AS dv1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY dim1

SELECT dim1 AS dv1
FROM dv1
WHERE `timestamp` >= ‘2022-01-01’ AND `timestamp` < ‘2022-01-02’
GROUP BY dim1
ORDER BY SUM(metric1)
LIMIT 15

SELECT SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’

SELECT dim1, dim2, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY dim1, dim2

SELECT dim1, dim2, SUM(metric1) AS m1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY 1, 2
ORDER BY 1, 2

SELECT DISTINCT dim1, dim2
FROM dv1

SELECT dim1, m1
FROM (
 SELECT dim1, SUM(metric1) AS m1
 FROM dv1
 WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
 GROUP BY dim1
)
WHERE dim1 in (‘A’, ‘B’)

SELECT key, SUM(m1) AS total
FROM (
 SELECT dim1 AS key, SUM(metric1) AS m1
 FROM dv1
 WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
 GROUP BY dim1

 UNION

 SELECT dim2 AS key, SUM(m1) AS m1
 FROM dv2
 WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
 GROUP BY dim2
GROUP BY key
ORDER BY total

SELECT rows.dim1, SUM(rows.m1) AS total
FROM (
 SELECT _.dim1,_.m1
 FROM (
 SELECT * FROM dv1
 WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
 ) _
 WHERE _.dim1 in (‘A’, ‘B’, ‘C’)
) rows
GROUP BY 1
ORDER BY total

WITH rows AS (
 WITH _ AS (
 SELECT * FROM data_ares
 WHERE `timestamp` BETWEEN ‘2021-01-01’ AND ‘2021-02-01’
 )
 SELECT _.item, _.units FROM _
 WHERE _.item IS NOT NULL
)
SELECT rows.item, SUM(rows.units) AS units
FROM rows WHERE rows.item in (‘A’, ‘B’, ‘C’)
GROUP BY rows.item

SELECT SUM(metric1) AS m1, dim1
FROM dv1
WHERE `timestamp` BETWEEN ‘2022-01-01’ AND ‘2022-01-02’
GROUP BY 2

Dimensions

You can select any of the dimensions available by default or defined in the data view. You select a dimension by its ID.

Metrics

The metrics available to select are:

You select a metric by its ID wrapped in a SUM(metric) expression just like you would do with other SQL sources.

You can use:

Counting distinct values

Due to the underlying nature of how Customer Journey Analytics works, the only dimension you can get an exact distinct count for is the adobe_personid dimension. The following SQL statements SELECT COUNT(DISTINCT adobe_personid) or SELECT APPROX_COUNT_DISTINCT(adobe_personid) return the value of the default persons metric, which is the count of distinct people. For other dimensions, an approximate distinct count is returned.

Conditional metrics

You can embed an IF or CASE clause in the SUM or COUNT functions to add additional filtering that is specific to a selected metric. Adding these clauses is similar to applying a filter to a metric column in a Workspace report table.

Examples:

SUM(IF(dim1 = 'X' AND dim2 = 'A', metric1, 0)) AS m1

SUM(CASE WHEN dim1 = 'X' AND dim2 = 'A' THEN metric1 END) AS m1

Inline calculations

You can apply additional math to metric expressions in your SELECT. This math can be used instead of defining the math in a calculated metric. The following table lists what types of expressions are supported.

Special columns

Timestamp

The timestamp special column is used to provide the date ranges for the query. A date range can be defined with a BETWEEN expression or a pair of timestamp >, >=, <, <= checks ANDed together.
The timestamp is optional and if no full range is provided, defaults are used:

The timestamp range is converted to a date range global filter in the RankedRequest.
The timestamp field can also be used in date/time functions to parse or truncate the event timestamp.

Date range

The daterange special column works similar to timestamp; however the filtering is limited to full days. The daterange is also optional and has the same range defaults as timestamp.
The daterange field can also be used in date/time functions to parse or truncate the event date.

The daterangeName special column can be used to filter your query using a named date range like Last Quarter.

Filter ID

The filterId special column is optional and is used to apply an externally defined filter to the query. Applying an externally defined filter to a query is similar to dragging a filter on a panel in Workspace. Multiple filter IDs can be used by AND-ing them.

Along with filterId, you can use filterName to use a filter’s name instead of ID.

Where clause

The WHERE clause is handled in three steps:

The handling is done by parsing the first level of ANDs in the WHERE clause. Each top-level AND-ed expression must match one of the above. Anything deeper than the first level of ANDs, or, if the WHERE clause uses ORs at the top level, is handled as an ad-hoc filter.

Sorting order

By default, the query sorts the results by the first selected metric in descending order. You can overwrite the default sorting order by specifying ORDER BY ... ASC or ORDER BY ... DESC. If you use ORDER BY, you must specify ORDER BY on the first selected metric.

You can also flip the order by using - (minus) in front of the metric. Both statements below result in the same ordering:

ORDER BY metric1 ASC

ORDER BY -metric1 DESC

General function support

Dimension function support

These functions can be used on dimensions in the SELECT, WHERE clause, or in conditional metrics.

String functions

Date-time functions

Partial Support

Some SQL functionality is only partially supported with the BI extension and does not return the same results you see with other databases. This specific functionality is used in SQL generated by various BI tools, for which the BI extension does not have an exact match. As a result, the BI extension focuses on a limited implementation that covers the minimum BI tool usage without throwing errors. See the table below for more details.