Client Analytics Manual

This manual explains how to use VectorVue analytics in commercial delivery and stakeholder communication.

Scope

Analytics capabilities include:

• security effectiveness scoring
• residual risk estimation
• detection coverage insight
• anomaly visibility
• explainable model output and confidence indicators

Portal Workflow

1. Open Overview and verify KPI and analytics cards load.
2. Open Analytics and review score, confidence, explanations, model version, and generation time.
3. Run simulation to evaluate proposed defensive improvements.
4. Open Risk to communicate trend and severity posture.

API Workflow

Use tenant-authenticated token and call:

• GET /ml/client/security-score
• GET /ml/client/risk
• GET /ml/client/detection-gaps
• GET /ml/client/anomalies
• POST /ml/client/simulate

Expected response contract:

• score
• confidence
• explanation
• model_version
• generated_at

Quality and Governance Rules

• Tenant isolation is mandatory for all analytics data paths.
• Predictions must reference model version and dataset lineage.
• Explanations must be available for client-facing decisions.
• Model promotion is explicit and controlled.

Troubleshooting

1. Verify migrations and seed data:

make deploy
make seed-clients

1. Verify workers are healthy:

docker compose ps

1. Validate endpoint behavior with API smoke and direct calls.