Contributor activity criteria

contributor-criteria.md

Contributor Activity Criteria

1. Data Collection Phase

Time Window: Analyzes the last N days (default 365, user-configurable 1-365)

Data Sources:

• Recent commits within the analysis window
• Recent issues and pull requests within the analysis window
• Review comments and participants

2. Bot Filtering

Bot Detection: Filters out automated accounts using multiple criteria:

• Username contains: "bot", "automated", "auto", "ci", "cd", "deploy", "build", "github-actions", "dependabot", "renovate", "codecov", "travis", "jenkins", etc.
• Display name contains similar bot indicators
• Email patterns suggesting automation

Result: Only human contributors are included in the analysis

3. Contributor Activity Tracking

For each human contributor, the system tracks:

• Commits: Number of commits authored
• Issues Created: Issues opened by the contributor
• PRs Created: Pull requests opened by the contributor
• Reviews Given: Code reviews provided to other PRs
• Comments Made: Estimated comments in issues/PRs (distributed among participants)

4. Email Classification

Each contributor's email is classified as:

• Company: Known corporate domains (microsoft.com, google.com, etc.)
• Personal: Gmail, Yahoo, Outlook, etc.
• Academic: .edu, .ac.uk, .edu.au domains
• Custom: User-provided company domains to filter
• "no email available": When email is missing or invalid

5. Activity Trend Analysis

Quarterly Breakdown: Divides analysis window into 4 quarters

Trend Calculation: Compares recent half (Q2+Q3) vs older half (Q0+Q1)

Trend Categories:

• Increasing: Recent activity > 1.5x older activity
• Decreasing: Recent activity < 0.67x older activity
• Stable: Activity levels roughly equal
• Insufficient Data: Less than 10 total activities

6. Sentiment Analysis

Eligibility: Only contributors with 10+ activities get sentiment analysis

Data Sources: Commit messages, issue comments, PR comments, review comments

Output: Average polarity, subjectivity, and sentiment label (positive/negative/neutral)

7. Final Output Structure

Each contributor entry includes all tracked metrics and classifications.

8. Sorting & Risk Analysis

Sorting: Contributors sorted by total activity (highest first)

Concentration Risk: Calculated based on top contributor's percentage of total activity

Activity Distribution: Shows how activity is distributed among top 1, 3, and 5 contributors

This comprehensive approach provides a detailed view of who's actively contributing to a repository, their engagement patterns, communication sentiment, and potential organizational affiliation based on email domains.