We are pleased to announce recent improvements to Healthcare Providers, B2C Customers, and B2B Customers data products, as well as new curation features for source records.

Healthcare Providers Improvements

A new option in Healthchcare Providers configuration provides you with more control over clustering.

You can now choose whether to use NPPES match data to inform clustering. For existing data products, this option is enabled automatically.

When you disable this option, Tamr uses only the NPI number to inform clustering (not any other data returned by NPPES).

Tamr continues to use high confidence NPI numbers to inform clustering, which are NPI numbers provided in source records and NPI matches based on the combination of the providers's name and address (nppes_match_status is either MATCH_NP or MATCH_NAME_STREET_ADDRESS).

See Healthcare Providers Data Product for more information.

B2C Customers Improvements

We have improved clustering to identify matches for records in which the national ID, data of birth, and country values are all null.

B2B Customers Improvements

If you are using Dun & Bradstreet enrichment, you'll notice that the Enhanced Source Records output now includes attributes provided by the DUNS Match service. These attributes are available in new and existing publish configurations. For backwards compatibility with previous published output, these attributes are not selected by default in existing configurations.

See DUNS Match Insight for the attributes provided by DUNS Match service.

Curation Options for Source Records

You can now perform curation actions from the Source Records page for your data product.

From the Source Records Actions menu, you can:

• Create a new entity from selected source records.
• Move records between clusters. When you select this option, the clusters associated with the selected records open in the Manage Cluster Details page, where you can compare the clusters and move records as needed.

Bug Fixes and Other Improvements

• Resolved an issue in which publishing datasets to Delta Lake or Parquet failed if the dataset was previously published to the same location.
• Resolved an issue in which changes made to Source Records table views were not saved correctly.
• General performance, styling, and usability improvements.