Why This Matters

This capstone builds a real dashboard that leadership uses for Copilot oversight. Combines API (PR-07 to PR-14), observability (PR-33 to PR-34), cost (PR-43).

Concept Explained

Dashboard pulls: seats (current, orphaned), usage (per user, team, surface), cost (credits, burn rate, forecast). Visualize. Alert.

How It Works

Build service: scheduled jobs pull via API, store in DB, web UI visualizes, alerts on anomalies. Present to leadership weekly.

Step-by-Step Tutorial

1. Pull data

2. Store

3. Visualize

4. Alert

5. Present

Real-World Example

A team built this dashboard. Leadership reviewed weekly. Caught over-users, orphaned seats, policy drift. Saved $500/month. Dashboard became essential leadership tool.

Example Prompts / Commands / Code

"""1. Overview: adoption rate, total cost, trends
2. Seats: current, orphaned, changes
3. Usage: per user, team, surface, model
4. Cost: credits, burn rate, forecast, per-team
5. Anomalies: over-budget, drift, orphans
6. Reports: weekly, monthly, quarterly
"""

"""- Backend: Python (FastAPI) or Node (Express)
- DB: PostgreSQL (time-series via TimescaleDB)
- Frontend: React + Chart.js or similar
- Alerts: Slack webhooks, email
- Auth: SSO (Okta, Google)
- Deploy: Cloud (AWS/Azure) or on-prem
"""

Common Mistakes

• Pulling too frequently — rate limits.
• No alerts — data without action.
• Not presenting to leadership — dashboard unused.
• No historical trends — can't see patterns.

Best Practices

• Scheduled pulls (daily seats, weekly usage).
• Time-series DB for trends.
• Visualize: overview, seats, usage, cost, anomalies.
• Alert on anomalies.
• Present to leadership weekly.

Troubleshooting

Practical Exercise

Professional Challenge

Frequently Asked Questions

Further Reading

• Metrics API

Official References

• Metrics API

Related lessons: PR-14, PR-47