What's new?
- Users can identify one or more benchmark fields in the Measure Metadata panel of measure workshop, from among the published fields of a measure.
- If any benchmark fields are identified, these will unlock two new perspectives in measure explorer: benchmark value and benchmark ratio. The benchmark ratio represents the measure value divided by the benchmark value.
- If benchmarks are defined on a measure, perspective tabs for the benchmark value and the benchmark ratio will appear at the bottom of the perspective tab group. If there are multiple benchmarks defined for a measure, users can cycle through each benchmark field by re-clicking on these perspective tabs.
- Because the benchmark values may vary across rows within the selected splits, the min and max benchmark value will be calculated, along with the average, for all the splits under review. Unless these values are all identical, the min and the max will be shown as an error bar in the measure explorer chart.
- Add new Scatter Plot screen to Analytics Portal, as a peer to Measure Explorer. Unlike the charts in Measure Explorer, scatter plots support up to three different measure expressions. The first is laid across the X axis, the second is laid across the Y axis, and the optional third measure expression drives the color of the bubble. Each split value is represented by a bubble, and the size of the bubble is determined by the denominator total. Users can click on each bubble to "Keep Only", "Exclude", or "View Cases" of the data represented by the bubble.
- Support ALTERNATE_S3_REGION environment variable, for dealing with S3 buckets in different regions than the Ursa Studio application. The value of the environment variable must be in the format us-east-2:my-bucket-1,my-bucket-2.
- Support keypair authentication for dedicated analytics-portal-only service account for Snowflake.
What improvements have been made to existing features?
- Support incremental load on single stacks, even those with multiple upstream objects, so long as only one of the single stack's upstream objects support incremental load.
- Support "Only Insert New Records" option for incremental load. This setting prevents any changes to existing records by allowing only the insertion of new records (i.e., without updating existing records, even when a new version of that record appears in the upstream object).
- Add row count to Last ELT column of Object Workshop dashboard.
- Allow users to load historical ELTs into preview, so as to run again or persist as a Saved ELT.
- When import objects with "Prevent Duplicate Imports" are added to an ETL under the "Run This and Downstream" instruction, these objects will be skipped if there are no actionable files to import, along with any downstream objects that are in the ELT on account of being in their downstream.
- Support a "28P" version option in the HCC model that will run the v28 model with the newly proposed 2027 coefficients as well as the proposed 2027 normalization factor, even when the payment dates are in previous years.
What's been fixed?
- Allow split on snapshot date for data review of objects.