Data snapshots and multiple breakdowns
Data snapshots in Platform Analytics allows for multiple breakdowns while analyzing your indicators (KPIs). This architecture uses a change data capture (CDC) process, which captures data changes from configurable tables that are optimized for generating scores and time series at run-time.
Classic Performance Analytics has always calculated multiple breakdowns applied to an indicator through a matrix. Mathematically, the number of calculations increases geometrically with the number of breakdowns. Practically, this approach limits you to a maximum of two breakdowns that can be applied to an indicator simultaneously.
Data snapshots does not use breakdown matrices. Thanks to the use of CDC instead, you can apply multiple breakdown levels. The ability to apply more than two levels of breakdown addresses one of the longstanding pain points of Performance Analytics. This feature deepens your ability to analyze data and broadens the insights you can gain through Performance Analytics.
- Data snapshots requires that your instance uses the RaptorDB Professional database.
- Domain-separated instances are not supported.
- To get data snapshots and enable multiple levels of breakdown, activate the Data Snapshots (com.snc.pa.mlb) plugin. A Performance Analytics subscription and the admin role are required. Consider testing data snapshots in a development environment before activating them on a production instance.
- When you have activated the Data Snapshots plugin, you still have to enable it for supported indicators. For more information, see Activate data snapshots.
- Only Platform Analytics experience components (Data visualizations, KPI Details) support data snapshots. You cannot show more than two levels of breakdowns in Core UI components like PA widgets and the Analytics Hub.
- Data snapshots collection jobs may result in increased storage use by Performance Analytics. The jobs copy subsets of the source tables and store every daily change for related records.
- Automated indicators that do not use scripted aggregations (non-scripted indicators).
- Formula indicators with only non-scripted, automated contributing indicators
- Formula indicators that contain other formula indicators, provided that ultimately the underlying indicators are non-scripted, automated indicators.
If an indicator does not support data snapshots, the reasons appear in a warning on the indicator's record page.