How can you determine when a project is done using the Machine Classification against Coding Values widget?

To determine when a project is done using the Machine Classification against Coding Values widget, identifying the presence of few or no purple bars (representing documents that are 'Not Reviewed') under the 50 mark for relevance is key. This threshold signifies that the majority of the documents have received attention and have been sorted appropriately.

When there are few or no 'Not Reviewed' documents below this relevance score, it typically indicates that the project has reached a level of completeness in the classification process. Most of the content has been classified as either relevant or non-relevant, allowing for a reliable assessment of the data. This gives confidence that the machine-learning model has categorized a majority of the documents adequately, and any remaining documents do not require extensive further review.

In contrast, if there are many purple bars over the 50 mark, it suggests that numerous documents haven’t been adequately reviewed, meaning the project is still ongoing and requires further attention. The total number of documents coded relevant and comments from reviewers provide useful insights but do not directly indicate completion in the same way that assessing the 'Not Reviewed' document distribution does.