Skip to content

Learning Graph Quality Metrics Report

Overview

  • Total Concepts: 561
  • Foundational Concepts (no prerequisites, other concepts depend on them): 3
  • Terminal Nodes (nothing depends on them, but have prerequisites): 96
  • Orphaned Nodes (completely disconnected, no edges): 0
  • Concepts with Dependencies: 558
  • Average Dependencies per Concept: 1.48

Graph Structure Validation

  • Valid DAG Structure: ✅ Yes
  • Self-Dependencies: None detected ✅
  • Cycles Detected: 0

Foundational Concepts

These concepts have no prerequisites:

  • 1: Artificial Intelligence
  • 46: Token
  • 182: Return On Investment

Dependency Chain Analysis

  • Maximum Dependency Chain Length: 98

Longest Learning Path:

  1. Artificial Intelligence (ID: 1)
  2. Machine Learning (ID: 2)
  3. Deep Learning (ID: 3)
  4. Neural Network (ID: 4)
  5. Generative AI (ID: 5)
  6. Large Language Model (ID: 6)
  7. Foundation Model (ID: 7)
  8. Transformer Architecture (ID: 8)
  9. Self-Attention (ID: 10)
  10. Pretraining (ID: 11)
  11. Fine-Tuning (ID: 12)
  12. Instruction Tuning (ID: 13)
  13. RLHF Alignment (ID: 14)
  14. Inference (ID: 15)
  15. Training Vs Inference Cost (ID: 16)
  16. Model Parameters (ID: 17)
  17. Model Parameter Count (ID: 18)
  18. Multimodal Model (ID: 20)
  19. Open-Source Model (ID: 21)
  20. Proprietary Model (ID: 22)
  21. Model Provider (ID: 23)
  22. API Endpoint (ID: 24)
  23. Prompt (ID: 25)
  24. Prompt Engineering (ID: 27)
  25. Zero-Shot Prompting (ID: 28)
  26. Few-Shot Prompting (ID: 29)
  27. Chain-Of-Thought Prompting (ID: 30)
  28. Retrieval-Augmented Generation (ID: 31)
  29. Embedding (ID: 32)
  30. Vector Database (ID: 33)
  31. Model Hallucination (ID: 34)
  32. Model Latency (ID: 35)
  33. Model Throughput (ID: 36)
  34. Batch Inference (ID: 37)
  35. Streaming Response (ID: 38)
  36. Model Versioning (ID: 39)
  37. Model Deployment (ID: 40)
  38. GPU Compute (ID: 92)
  39. CPU Inference (ID: 93)
  40. Compute Cost (ID: 94)
  41. Cloud Compute Pricing (ID: 95)
  42. Reserved Capacity Pricing (ID: 97)
  43. Spot Instance Pricing (ID: 98)
  44. Dedicated Endpoint (ID: 100)
  45. Multi-Tenant Endpoint (ID: 101)
  46. Autoscaling (ID: 102)
  47. Warm Pool (ID: 104)
  48. Compute Utilization Rate (ID: 105)
  49. Idle Capacity Cost (ID: 106)
  50. Data Transfer Cost (ID: 107)
  51. Storage Cost (ID: 108)
  52. Vector Store Cost (ID: 109)
  53. Network Egress Cost (ID: 110)
  54. Load Balancing (ID: 111)
  55. Model Serving Framework (ID: 112)
  56. Inference Server (ID: 113)
  57. GPU Memory Constraint (ID: 115)
  58. Concurrency Limit (ID: 117)
  59. Throughput-Latency Tradeoff (ID: 118)
  60. Peak Load Provisioning (ID: 119)
  61. Capacity Planning (ID: 120)
  62. Infrastructure As Code (ID: 121)
  63. Multi-Region Deployment (ID: 122)
  64. Edge Deployment (ID: 123)
  65. Hybrid Cloud Strategy (ID: 124)
  66. Total Compute Footprint (ID: 125)
  67. Carbon Cost Of Compute (ID: 126)
  68. Energy Cost Per Query (ID: 127)
  69. Infrastructure Monitoring (ID: 128)
  70. Cost Alerting (ID: 129)
  71. Real-Time Cost Monitoring (ID: 276)
  72. Spend Forecasting (ID: 278)
  73. Budget Variance Analysis (ID: 279)
  74. Cost Data Warehouse (ID: 280)
  75. Data Pipeline For Cost Analytics (ID: 281)
  76. ETL For Usage Data (ID: 282)
  77. Data Quality For Cost Metrics (ID: 283)
  78. Sampling Strategy For Telemetry (ID: 284)
  79. Data Retention Policy (ID: 285)
  80. Multi-Tenant Cost Attribution (ID: 286)
  81. Departmental Cost Rollup (ID: 287)
  82. Project-Level Cost Tracking (ID: 288)
  83. Per-User Cost Tracking (ID: 289)
  84. Per-Feature Cost Tracking (ID: 290)
  85. Vendor Invoice Reconciliation (ID: 291)
  86. Billing API Integration (ID: 292)
  87. Cost Data Normalization (ID: 293)
  88. Cross-Provider Cost Aggregation (ID: 294)
  89. Dashboard Refresh Cadence (ID: 296)
  90. Data Governance For Metrics (ID: 297)
  91. Audit Trail For Spend (ID: 298)
  92. Metric Definition Documentation (ID: 299)
  93. Single Source Of Truth (ID: 300)
  94. Instrumentation Maturity Model (ID: 301)
  95. Cost Observability Roadmap (ID: 302)
  96. Real-Time Alert Fatigue (ID: 304)
  97. Data Lineage For Cost Metrics (ID: 305)
  98. Instrumentation Ownership Model (ID: 306)

Terminal Nodes Analysis

Terminal nodes are concepts that nothing else depends on but have prerequisites. They represent natural endpoints of learning paths — culminating or specialized concepts.

  • Total Terminal Nodes: 96 (17.1% of all concepts)
  • Healthy Range: 5-40% of total concepts

Concepts at the end of learning paths:

  • 9: Attention Mechanism
  • 19: Context Window
  • 26: System Prompt
  • 52: Output Token
  • 62: Blended Token Rate
  • 67: Token Consumption Rate
  • 69: Prompt Length
  • 73: Long-Context Pricing
  • 77: Semantic Caching
  • 80: Chunking Strategy
  • 88: Token Cost Benchmark
  • 90: Token Efficiency Score
  • 91: Tokenizer Version Drift
  • 96: On-Demand Pricing
  • 99: Serverless Inference
  • 103: Cold Start Latency
  • 114: Hardware Accelerator
  • 116: Batch Size Tuning
  • 131: Infrastructure Cost Dashboard
  • 132: Prompt Optimization

...and 76 more

Orphaned Nodes Analysis

Orphaned nodes are completely disconnected concepts with no inbound AND no outbound edges. These indicate a quality problem — every concept should connect to the graph.

  • Total Orphaned Nodes: 0

✅ No orphaned nodes detected. All concepts are connected to the graph.

Connected Components

  • Number of Connected Components: 1

✅ All concepts are connected in a single graph.

Indegree Analysis

Top 10 concepts that are prerequisites for the most other concepts:

Rank Concept ID Concept Label Indegree
1 227 GenAI ROI Metric 9
2 183 Total Cost Of Ownership 7
3 6 Large Language Model 5
4 94 Compute Cost 5
5 95 Cloud Compute Pricing 5
6 156 Summarization Preprocessing 5
7 437 AI Governance Framework 5
8 5 Generative AI 4
9 23 Model Provider 4
10 27 Prompt Engineering 4

Outdegree Distribution

Dependencies Number of Concepts
0 3
1 320
2 215
3 18
4 5

Recommendations

  • ✅ Terminal node percentage (17.1%): Within healthy range (5-40%)
  • ✅ DAG structure verified: Graph supports valid learning progressions
  • ℹ️ Long dependency chains (98): Ensure students can follow extended learning paths
  • ℹ️ Consider adding cross-dependencies: More connections could create richer learning pathways

Report generated by learning-graph-reports/analyze_graph.py