Skip to content

READABLE_ANALYTICS.md

Latest commit

 

History

History
220 lines (172 loc) · 7.52 KB

READABLE_ANALYTICS.md

File metadata and controls

220 lines (172 loc) · 7.52 KB

Comprehensive Backup Analytics - Human-Readable Output

What Changed

Instead of showing raw metrics like this:

AI-Assisted Commits: 234
Solo Commits: 156
Velocity with AI: 2.34 commits/hour
Velocity solo: 1.89 commits/hour

We now show actionable insights like this:

Example Output

═══════════════════════════════════════════════════
           📊 YOUR PRODUCTIVITY ANALYSIS
═══════════════════════════════════════════════════

🎯 Overall Productivity Score: 67/100 (Grade: C+)

  Breakdown:
    • AI Effectiveness:   78/100  (40% weight)
    • Shell Efficiency:   58/100  (30% weight)
    • Workflow Quality:   64/100  (30% weight)

🤖 AI Impact on Productivity
    • AI-Assisted Commits: 234 (60.0%)
    • Velocity Improvement: +23.8%
    • Code Volume: 12,456 lines with AI (67.3%)
    • Copy-Paste Incidents: 12

🐚 Shell Command Analysis
    • Total Commands: 8,945
    • Failure Rate: 26.8%
    • Time Wasted: 42.3 hours
    • Struggle Sessions: 89
    • Productivity Score: 58/100

🔗 Workflow Correlation Analysis
    • Full Cycle Workflows: 67 (Struggle → AI → Commit)
    • AI Helpfulness Rate: 75.3%
    • Shell → AI: 89 instances
    • AI → Commit: 145 instances

═══════════════════════════════════════════════════
           🎯 ACTIONABLE RECOMMENDATIONS
═══════════════════════════════════════════════════

🟠 HIGH - Shell Efficiency
  Issue: High command failure rate: 26.8%
  Action: Use Ctrl+R for history, create aliases, ask AI earlier
  Impact: Save ~42.3 hours/month

🟠 HIGH - Code Quality
  Issue: Detected 12 copy-paste incidents from Claude
  Action: Take time to understand code before committing. Ask Claude to explain complex parts.
  Impact: Reduce bugs by 27%, improve code understanding

🟡 MEDIUM - Workflow
  Issue: Detected 89 struggle sessions (multiple retries)
  Action: Ask Claude earlier when stuck. Average 4+ retries before AI help - ask sooner!
  Impact: Reduce frustration, solve problems 3x faster

🟡 MEDIUM - AI Effectiveness
  Issue: AI only resolves 75.3% of struggles
  Action: Provide more context when asking Claude. Include error messages, relevant code, and what you've tried.
  Impact: Increase AI success rate to 90%+

═══════════════════════════════════════════════════

  ✓ Full analysis saved: /tmp/comprehensive-analytics.json

New Analytics Modules

1. Shell Productivity Analyzer (src/shell_analytics.rs)

What it analyzes:

  • Command success/failure rates
  • Time wasted on failed commands
  • Struggle sessions (3+ retries on related commands)
  • Common error patterns
  • Productivity scoring (0-100)

Key insights:

  • "You have a 26.8% command failure rate - that's wasting 42 hours!"
  • "You struggle with build failures (89 sessions detected)"
  • "Most failed command: npm install (234 failures)"

2. Workflow Correlation Engine (src/workflow_correlation.rs)

What it tracks:

  • Shell Error → Claude Help → Commit patterns
  • Time to resolution with/without AI
  • AI helpfulness rate (% of struggles resolved)
  • 6 workflow pattern types

Key insights:

  • "You ask Claude after 4 failed attempts on average - ask sooner!"
  • "AI solves 75% of your struggle sessions"
  • "Git conflicts take 12 minutes with AI vs 45 minutes without"

3. Comprehensive Analytics Engine (src/comprehensive_backup_analytics.rs)

What it does:

  • Integrates AI Impact + Shell Analytics + Workflow Correlation
  • Generates actionable recommendations with priority
  • Calculates overall productivity score (0-100 with A+ to F grades)

Scoring breakdown:

  • 40% AI Effectiveness - How well you use AI (velocity + code quality)
  • 30% Shell Efficiency - How productive you are in the terminal
  • 30% Workflow Quality - How well your tools work together

How to Use

# Standard backup with comprehensive analytics
./target/release/vibedev backup \
  --output ~/backups \
  --include-git \
  --include-history \
  --analyze-impact

# This will:
# 1. Backup AI logs (Claude, Cursor, etc.)
# 2. Backup git commit history from all repos
# 3. Backup shell history (sanitized from API keys)
# 4. Generate comprehensive productivity analysis
# 5. Show human-readable recommendations

# Just the analytics (no backup)
./target/release/vibedev backup \
  --analyze-impact \
  --include-git \
  --include-history \
  --output /tmp

What Gets Analyzed

AI Impact Analysis

  • ✅ Correlates Claude conversations with git commits
  • ✅ Measures velocity improvement (commits/hour with vs without AI)
  • ✅ Detects collaboration patterns
  • ✅ Identifies copy-paste incidents (suspiciously fast commits)
  • ✅ Tracks learning curve over time

Shell Productivity

  • ✅ Parses shell history (.bash_history, .zsh_history)
  • ✅ Detects failed commands using error patterns
  • ✅ Identifies struggle sessions (multiple retries)
  • ✅ Calculates time wasted on errors
  • ✅ Productivity scoring

Workflow Correlation

  • ✅ Detects Shell → AI → Commit patterns
  • ✅ Measures AI helpfulness (% of struggles resolved)
  • ✅ Tracks time to resolution
  • ✅ Identifies workflow anti-patterns

Recommendations System

Recommendations are priority-ranked and actionable:

Priority Emoji When Used
🔴 CRITICAL Red Productivity score < 60, severe issues
🟠 HIGH Orange Failure rate > 20%, copy-paste > 20 incidents
🟡 MEDIUM Yellow Struggle sessions > 50, AI helpfulness < 50%
🟢 LOW Green Positive reinforcement, velocity > 30%

Each recommendation includes:

  • Issue - What's wrong (with metrics)
  • Action - Specific steps to improve
  • Impact - Expected improvement (hours saved, % increase)

Files Created

After running with --analyze-impact, you get:

~/backups/
  comprehensive-analytics-20260105-085500.json  # Full JSON report
  ai-logs-20260105-085500.tar.gz                 # AI tool logs
  git-logs-20260105-085500.tar.gz                # Git commit history
  shell-history-20260105-085500.tar.gz           # Sanitized shell history

Integration Status

✅ Module imports added to src/main.rs ✅ Comprehensive analytics integrated in backup command ✅ Human-readable output with emojis and formatting ✅ Actionable recommendations with priority ranking ✅ Productivity scoring with A-F grades ✅ Compilation successful (all errors fixed) ⏳ Testing - Ready for real-world use

Technical Details

Data sources:

  • Claude conversations (~/.claude/projects/*/history.jsonl)
  • Git commits (git log --numstat from all repos in home dir)
  • Shell history (~/.bash_history, ~/.zsh_history, etc.)

Analysis pipeline:

  1. AIImpactAnalyzer loads Claude + Git data
  2. ShellAnalyzer parses shell history
  3. WorkflowAnalyzer correlates across tools
  4. ComprehensiveAnalyticsEngine generates recommendations
  5. Output displayed with colored, formatted text

Privacy:

  • Shell history is sanitized (20+ regex patterns for API keys, passwords, etc.)
  • All data stays local (no network calls)
  • JSON reports can be reviewed before sharing