Peak Hours & Commit Types

Peak Hours

Peak hours analysis shows when a contributor is most active, based on the hour each commit was authored.

How It Works

1. Extract the hour (0–23) from each commit's committed_at timestamp
2. Build a histogram counting commits per hour across the date range
3. Identify peak hours (highest-frequency bins)

What Peak Hours Reveal

Pattern	Possible Interpretation
Morning cluster (8–12)	Focused morning coder, meetings in afternoon
Afternoon cluster (13–17)	Morning meetings, codes after lunch
Evening cluster (18–22)	After-hours worker or different timezone
Late night (22–4)	Night owl, timezone mismatch, or crunch
Bimodal (morning + evening)	Meetings midday, codes around them

WARNING

Peak hours use commit timestamps, which may not reflect actual coding time. A developer might code for hours and commit once. Use this as directional, not precise.

Commit Types

Commit type analysis breaks down a contributor's work by conventional commit categories.

Supported Types

Type	Description
feat	New feature or functionality
fix	Bug fix
refactor	Code restructuring without behavior change
test	Adding or updating tests
docs	Documentation changes
chore	Maintenance, tooling, config
style	Code style/formatting changes
perf	Performance improvements

Metrics

For each contributor:

• Count per type
• Percentage per type (out of total commits)

Healthy Distributions

There's no single "right" distribution, but some patterns are informative:

Pattern	Signal
High feat ratio	Active feature development phase
High fix ratio (> 50%)	Bug-heavy period — triggers alert
Balanced feat/fix/refactor	Sustainable development cycle
Zero test commits	May indicate low test discipline — triggers alert
High chore ratio	Heavy tooling/maintenance phase

Test Ratio

The test ratio measures what percentage of a contributor's commits touch test-related files.

Detection

Files are considered test-related if their path contains:

• test
• spec
• __tests__

$$\text{test\_ratio}=\frac{\text{commits touching test files}}{\text{total commits}}\times 100$$

Thresholds

Ratio	Interpretation
> 30%	Strong test discipline
10–30%	Moderate test coverage habits
< 10%	Low test discipline — triggers alert (if ≥ 5 commits in period)