Dataset: Issues & Pull Requests
Entity: Pull Requests, Issues
Field ID: approval_count
Type: Integer
Description: The number of peer review approvals received on this pull request. For issues, the count is the sum of approvals from resolving pull requests.
Source: Calculated from pull request reviews
Transformation logic:
Pull requests: the number of pull request reviews in the
APPROVEDstate. For a given reviewer, only the latest non-discarded review is considered (previous reviews are ignored).Issues: the sum of approvals from resolving pull requests. Resolving pull requests are pull requests that reference the issue via auto-closing keywords.
From: |
|
Github (PRs, Issues) | Calculated |
Gitlab (PRs, Issues) | Calculated |
Bitbucket (PRs) | Calculated |
Azure DevOps (PRs, Issues) | Calculated Required and optional reviewers both participate in the calculation. |
JIRA (Issues) | Calculated from resolving PRs |
ClickUp (Issues) | Calculated from resolving PRs |
Trello (Issues) | Calculated from resolving PRs |
Reporting Use Cases
The approval_count field is valuable for creating metrics and reports that help you understand and improve your code review processes.
Here are some ways you can use it:
Filtering and Compliance: You can filter pull requests based on the number of approvals they have received. For instance, you can create a report that lists all merged pull requests that did not meet the minimum required approvals by using a filter like
approval_count < required_approval_count. This is a great way to audit and enforce your review policies.Performance and Engagement Metrics: Since approval_count is a number, you can use it in a variety of calculations and aggregations to measure team performance. For example:
You can calculate the average number of approvals per pull request using
AVG(approval_count)to gauge the overall engagement in the review process.You could track the
MAX(approval_count)to see which pull requests generated the most discussion and consensus.It can also be used to create a KPI that monitors the percentage of pull requests merged with at least one approval:
ROUND(COUNT_IF(approval_count > 0) / COUNT() * 100, 2).
Workload and Bottleneck Analysis: By combining approval_count with other dimensions, you can analyze review workload distribution. For example, a bar chart with author_username as a dimension and
AVG(approval_count)as a metric can highlight if certain developers' pull requests consistently require more reviews than others, which could be an indicator of complexity or a need for mentoring.