User GuidesAPI ReferenceRelease Notes
Doc HomeHelp CenterLog In
User Guides

Tasks for Curators and Reviewers in the Categorization Workflow

When deciding how to categorize records into a taxonomy, you can seek expertise from subject matter experts. In Tamr, subject matter experts act as Reviewers. Once reviewers assign labels, curators choose between them and verify them.

The following diagram illustrates the tasks for each user role in Tamr:

  • Administrators assign roles to users as reviewers and curators.
  • Reviewers create labels, and also upvote or downvote existing labels. They can also skip records.
  • Curators verify label assignments and help choose a single label from multiple suggested labels.
1382

Tasks for Tamr Administrators, Curators, and Reviewers.

Accepting or Rejecting Tamr Suggestions

Once Tamr has attempted to categorize the records, you can go through and approve items that are correct. Other users can upvote and downvote Tamr suggestions. You can also change categories for records that Tamr classified incorrectly. If the project has multiple Reviewers, they can upvote or downvote an assigned category to indicate agreement. A Curator then reviews the labels and chooses a winning categorization, known as the verified categorization. Tamr learns only from the verified categorizations.

1186

Reviewing feedback on categorizations.

Choosing One Categorization from Multiple Classifications

If the data experts, or reviewers, provide more than one categorization, you, as a curator, can choose between them by checking a categorization in the details tab. The categorization that was selected by the data expert is indicated with a green checkmark. In some cases, Tamr provides an alternative categorization which you can also choose to select or ignore. Categorization labels indicate who labeled them and when, who upvoted them, and which one is currently chosen. Tamr learns only from the chosen categorization, since a record can only be assigned to one category in the taxonomy.

704

The Details tab with multiple categories provided by data experts.