feat: Learning analytics dashboard

[github-actions[bot]]

Originally posted as issue #271 by @informatJonas
#271

Summary

Add an analytics dashboard to the web viewer showing trends, patterns, and insights about captured observations over time.

Problem

Users accumulate observations but have no visibility into:

• How their work patterns evolve
• Which topics they work on most
• Productivity trends over time
• Most valuable learnings

Proposed Solution

1. Dashboard Overview

Add a new "Analytics" tab in the web viewer with:

┌─────────────────────────────────────────────────────────────┐
│  📊 Learning Analytics                      Last 30 days ▼  │
├─────────────────────────────────────────────────────────────┤
│                                                             │
│  Observations This Period    Sessions    Avg per Session   │
│        ████ 142              ████ 28         5.1           │
│        +23% vs last period   +15%           +8%            │
│                                                             │
├─────────────────────────────────────────────────────────────┤
│  Activity Heatmap                                           │
│  ┌─────────────────────────────────────────────────────┐   │
│  │ Mon ▓▓░░▓▓▓▓░░░░▓▓▓▓▓▓░░░░░░░░                     │   │
│  │ Tue ░░▓▓▓▓▓▓▓▓░░░░▓▓▓▓░░░░░░░░                     │   │
│  │ Wed ▓▓▓▓▓▓░░░░▓▓▓▓▓▓▓▓▓▓░░░░░░                     │   │
│  │ ...                                                 │   │
│  └─────────────────────────────────────────────────────┘   │
│                                                             │
├─────────────────────────────────────────────────────────────┤
│  Top Concepts                  Top Observation Types        │
│  ┌────────────────────┐       ┌────────────────────┐       │
│  │ problem-solution ████ 34  │ │ bugfix      ████████ 45  │ │
│  │ how-it-works    ███ 28   │ │ feature     ██████ 38    │ │
│  │ pattern         ██ 19    │ │ refactor    ███ 22       │ │
│  │ gotcha          █ 12     │ │ decision    ██ 15        │ │
│  └────────────────────┘       └────────────────────┘       │
│                                                             │
├─────────────────────────────────────────────────────────────┤
│  Recent Highlights                                          │
│  • Most productive day: Tuesday (avg 8.2 observations)      │
│  • Longest session: 3.5 hours (Dec 10, feature work)        │
│  • Most active project: claude-mem (67 observations)        │
│  • Knowledge growth: +142 observations this month           │
│                                                             │
└─────────────────────────────────────────────────────────────┘

2. Metrics to Track

Activity Metrics:

• Observations per day/week/month
• Sessions per day/week
• Average observations per session
• Session duration (if trackable)

Content Metrics:

• Distribution by observation type (bugfix, feature, etc.)
• Distribution by concept (how-it-works, gotcha, etc.)
• Most frequently modified files
• Most active projects

Trend Metrics:

• Week-over-week changes
• Month-over-month growth
• Moving averages

3. Insights Engine

Generate automated insights:

• "You fixed 40% more bugs this week"
• "Most of your decisions are about API design"
• "Tuesday is your most productive day"
• "You've been focusing on authentication for 3 sessions"

4. Export & Sharing

• Export analytics as PDF/PNG for reports
• Weekly email digest (optional)
• Share specific insights

Implementation Notes

• All analytics computed from existing observation data
• No additional tracking needed
• Cached calculations for performance
• Date range selector (7d, 30d, 90d, all time)

Use Cases

1. Self-Reflection: Understand work patterns
2. Progress Tracking: Visualize learning over time
3. Reporting: Show productivity metrics
4. Optimization: Identify most productive times/topics

🤖 Generated with Claude Code