Dataset: Reviews from pull requests
Entity: Pull Request Review
Field ID: body
Type: Text
Description: The content of the review.
Source: App
From: |
|
Github (Reviews) |
|
Gitlab (MRs) |
|
Bitbucket (Reviews) | None |
Azure DevOps (PR Threads) | None |
Jira | N/A |
ClickUp | N/A |
Trello | N/A |
Reporting Use Cases
The Body field from the Reviews dataset contains the main summary comment a reviewer leaves when they submit their review. It is a rich source of qualitative feedback that can be analyzed to understand the nature and substance of your code review process.
Filtering for Specific Feedback: You can create reports that find reviews based on the content of the summary comment.
Find Simple Approvals: You can isolate reviews that are simple "looks good to me" approvals by using a filter like
Body ~ "(?i)LGTM|Approved".Identify Key Concerns: Search for reviews that mention critical topics like "security," "performance," or "deprecation" to audit how these concerns are being addressed in your review process, using a filter like
Body ~ "security".
Providing Context in Lists: In any drilldown or list of individual reviews, including the body as a dimension is essential. It shows the reviewer's overall conclusion and provides critical context alongside the review's state (e.g., a
CHANGES_REQUESTEDstate is much clearer when you can see the accompanying comment).Custom Formulas for Categorization: You can use custom formulas to classify reviews based on their content, turning unstructured text into a valuable dimension for analysis.
Substantive vs. Simple Reviews: You can create a dimension to distinguish between brief approvals and more detailed feedback with a formula like
IF(body ~ "(?i)lgtm|approved", "Simple Approval", "Substantive Feedback"). Using this in a pie chart can show you the ratio of in-depth reviews to simple sign-offs.Counting Question-Based Reviews: You can measure how often reviews include questions (which might indicate a need for clearer PR descriptions) with a metric like
COUNT_IF(body ~ "\\?").