Home/Docs/Insights

Insights Explained

Every insight rule, its trigger conditions, confidence scoring, and real-world examples.

What are Insights?

Insights are actionable intelligence derived from repository metrics. Unlike raw numbers, insights provide specific recommendations based on patterns in the data.

🎯

Rule-Based

Each insight triggered by specific metric thresholds

📊

Confidence Scored

0-1 rating indicating insight reliability

💡

Actionable

Specific recommendations for improvement

Insight Categories

📈

Activity

Insights about commit frequency, recency, and development velocity

⚡

Responsiveness

Insights about issue response times and PR merge velocity

📊

Growth

Insights about community adoption and popularity trends

🔧

Maintenance

Insights about issue resolution and project upkeep

🌍

Diversity

Insights about contributor distribution and community breadth

Activity Insights

Low Commit Activity

warning0.9 confidence
Condition:commitFrequency < 0.1 (less than 9 commits in 90 days)
Message:Low commit activity: only {commitsLast90Days} commits in the last 90 days
Evidence:Consistent low activity indicates the project may be abandoned or in maintenance mode
Example:A repository with only 3 commits in 90 days triggers this insight

High Commit Activity

info0.8 confidence
Condition:commitFrequency > 2 (more than 180 commits in 90 days)
Message:High commit activity: {commitsLast90Days} commits in the last 90 days
Evidence:High commit frequency suggests active development and regular maintenance
Example:A repository with 250 commits in 90 days shows strong development activity

Repository Inactive

warning0.95 confidence
Condition:lastCommitAge > 30 days
Message:Repository inactive: last commit was {lastCommitAge} days ago
Evidence:No recent commits indicate the project may no longer be actively maintained
Example:Last commit 45 days ago triggers this insight with high confidence

Responsiveness Insights

Slow Issue Response

warning0.85 confidence
Condition:issueResponseTime > 168 hours (7 days)
Message:Slow issue response: average response time is {issueResponseTime / 24} days
Evidence:Slow responses to issues can discourage potential contributors
Example:Average response time of 10 days indicates responsiveness issues

Slow PR Merge Velocity

warning0.8 confidence
Condition:prMergeVelocity > 14 days
Message:Slow PR merge velocity: average merge time is {prMergeVelocity} days
Evidence:Slow PR processing can frustrate contributors and slow development
Example:PRs taking an average of 21 days to merge

Growth Insights

Stable Popularity

info0.7 confidence
Condition:stars > 1000 AND starGrowthRate < 1
Message:Stable popularity: {stars} stars with steady growth
Evidence:Established projects with consistent star counts show stable community interest
Example:Repository with 1500 stars and minimal recent growth

Maintenance Insights

Low Issue Resolution Rate

warning0.8 confidence
Condition:issueResolutionRate < 0.5
Message:Low issue resolution rate: only {issueResolutionRate * 100}% of recent issues resolved
Evidence:Unresolved issues indicate maintenance backlog or resource constraints
Example:Only 30% of issues from the last 90 days have been resolved

Backlog of Old Issues

error0.9 confidence
Condition:openIssues > 50 AND averageIssueAge > 90 days
Message:Backlog of old issues: {openIssues} open issues with average age of {averageIssueAge} days
Evidence:Large numbers of old open issues suggest maintenance challenges
Example:75 open issues averaging 120 days old

High PR Acceptance Rate

info0.8 confidence
Condition:prAcceptanceRate > 0.8
Message:High PR acceptance rate: {prAcceptanceRate * 100}% of PRs are merged
Evidence:High acceptance rates indicate welcoming maintainer culture
Example:85% of submitted PRs are eventually merged

Diversity Insights

Limited Contributor Diversity

info0.75 confidence
Condition:contributorDiversity < 0.3
Message:Limited contributor diversity: most commits from few contributors
Evidence:Low diversity may indicate bus factor risks or limited community participation
Example:80% of commits from a single contributor

Strong Contributor Community

info0.85 confidence
Condition:contributors > 10 AND contributorDiversity > 0.5
Message:Strong contributor community: {contributors} contributors with good diversity
Evidence:Diverse contributor base indicates healthy, collaborative project
Example:25 contributors with high diversity ratio

Confidence Scoring

Every insight includes a confidence score (0-1) indicating how reliable the insight is based on data quality and pattern strength.

0.95
Very High - Strong evidence, low false positive rate
Repository inactive (last commit >30 days)
0.85-0.9
High - Clear patterns with reliable indicators
Slow issue response, strong contributor community
0.7-0.8
Medium - Good indicators but some ambiguity possible
High commit activity, stable popularity
0.75
Moderate - Useful signal but may have edge cases
Limited contributor diversity

How Insights Help Maintainers

Prioritization

Focus on the most critical issues first based on insight levels and confidence scores.

Actionable Feedback

Each insight provides specific evidence and context for making informed decisions.

Trend Monitoring

Track how insights change over time to measure the impact of maintenance efforts.

Community Communication

Use insights to communicate project health and priorities to contributors.

← Back to Documentation • Health Score Calculation →