Personalized Learning Paths

[ringo380]

Goal

Implement skill-based progression tracking and generate adaptive learning recommendations.

Proposed Improvements

• Track user query patterns and skill progression
• Generate personalized improvement recommendations
• Create topic-based knowledge assessments
• Build interactive SQL tutorials based on user queries

Success Metrics

• User skill level classification accuracy >85%
• Recommendation relevance rating >4/5
• Learning path completion rate >60%

Related Components

• analyzer/ml/recommendations/learning_paths.py (was learning_path_generator.py)
• analyzer/ml/recommendations/personalization_engine.py (was feedback_personalization.py)

[ringo380]

Reconciliation: learning-path + personalization modules exist and are wired (authenticated path)

Both modules this issue proposes already exist and are reachable; the gap is per-user data and validation, not implementation.

Stale paths -> real locations:

• analyzer/ml/learning_path_generator.py -> analyzer/ml/recommendations/learning_paths.py (LearningPathGenerator)
• analyzer/ml/feedback_personalization.py -> analyzer/ml/recommendations/personalization_engine.py (FeedbackPersonalizationEngine)

Reachability: analyzer/ml/analysis/unified_analyzer.py:31 imports LearningPathGenerator and :36 imports FeedbackPersonalizationEngine; UnifiedQueryAnalyzer.analyze_query() runs at analyzer/views/query_grading_views.py:216 on the authenticated path only (query_grading_views.py:199). Personalization inherently needs per-user history, and with effectively no authenticated traffic there is none to personalize on.

Suggested reframe: gate this behind real authenticated usage (and the feedback accumulation tracked in #92). The modules exist; the open work is exercising them with real user query history and validating the success metrics, not building the generators.