Understanding Operational Analytics

Direct Data Access

The analytics visualization product's canvases report directly off of the operational system (with the exception of batch run analytics) thus eliminating the need of an ETL process. Selected tables and views have been designated as dimensions and facts for the purpose of generating the Start Schemas used by these canvases.

Batch Run Analytics

There is a high volume of batch run data in the operational system that changes frequently and is recorded in a complex set of relationships. The product provides snapshot tables to capture only the batch run and thread data required for analytics. Using these tables for batch subject areas both improves performance and supports extracting the data for upload to other environments.

The snapshot tables are populated in increments. The Batch Run Analytics Snapshot Update and the Batch Thread Analytics Snapshot Update processes are run an ongoing basis; each snapshot update selects only the batch runs and threads not yet included in the snapshot. Parameters are supplied to restrict the selected data to a range of months, to support the initial uptake of data. It is recommended that batch run data is updated before the batch thread data, as only threads with corresponding records in the batch run snapshot will be added. Refer to batch controls F1BRANSN and F1BTANSN for more information.

The Batch Run Analytics Data Export and Batch Thread Analytics Data Export processes are provided to extract the batch data in increments for upload to other environments. These processes assume that a Data Export Control exists for each of the snapshot maintenance objects, to capture the date of the previous extract as a reference point for selecting the next set of records. Refer to batch controls F1BRANEX and F1BTANEX for more information.

Calendar and Time Dimensions

Analytic reports rely on calendar and time dimensions to support various hierarchical grouping by date and time.

The Calendar dimension provides a level-based definition of the standard calendar and fiscal calendar in a flattened representation that is commonly used in data warehouses. This is necessary to enabling customers to group by calendar week, months, quarter and fiscal calendar period, quarter and fiscal year etc. Note that fiscal information about a specific date is optional and sourced from your edge application specific accounting calendar

In the same way, the Time dimension provides a level-based definition of each minute in a day, supporting reports that group by hours, AM/PM etc.

Bucket Configuration

The analytics visualization product provides support for defining a set of ranges, each representing a bucket for which extracted measures can be grouped and classified under the relevant bucket. The framework product provides support for viewing and defining the buckets. Refer to Bucket Configuration for more information.

Analytics Attribute Mapping

The product supports mapping of characteristics to user defined fields associated with dimensions in the analytics visualization product. Each characteristic table that is associated with a dimension table is provided with pre-generated user defined fields identified by unique column sequence numbers in the analytics visualization product. The mapping of characteristics to dimension user defined fields is maintained directly on the dimension portal. You may also review these mappings using the Analytics Attribute Mapping portal.

Logical Dimension

The Element Name of a foreign key field to a dimension represents a distinct logical version of the dimension. This allows for different logical dimensions to be based on the same dimension table. For example, a "Location" dimension may be referenced as a "Main Location" on one fact and also as an "Alternate Location" on the same or other fact. While both elements refer to the same dimension as foreign keys, each represent a distinct logical dimension via their element names. In the same way, while each date field is implicitly considered as a foreign key to the Calendar dimension, each distinct element name establishes a unique logical dimension. For example, a "Start Date" element name and an "End Date" element name are distinct logical dimensions of the underlying Calendar dimension.

Some logical dimensions may represent an aggregated or summarized form of another logical dimension. They are also known as "shrunken" dimensions. For example, the "Calendar Month Dimension" view is an aggregated dimension of the Calendar dimension. Establishing a link between a summarized dimension and its detailed dimension allows the analytics visualization tool to drill from one fact to another for performance reasons when they are indirectly linked via the same logical dimension but at different aggregation levels. For example, "Fact A" represents daily records and "Fact B" aggregates the same measures at a monthly level. Associating the "month" field in "Fact B", which is a foreign key to the "Calendar Month Dimension" view, with the logical dimension name of the "date" field in "Fact A" allows analytic queries to use "Fact B" for queries at the monthly level or above which perform better than performing them on the more granular "Fact A".

Analytics Configuration

Depending on which product or products you have installed, there may be some configuration needed for the analytics visualization product. Refer to Defining Analytics Options for more information.