Home > POC Agenda > Instructor Guide

👩‍🏫 Instructor Guide - Casino Fabric POC¶

Last Updated: 2026-04-15 | Version: 2.0 Status: ✅ Final | Maintainer: Documentation Team

🏠 Home > 📆 POC Agenda > 📘 Instructor Guide

Version: 1.0 Last Updated: 2026-01-21 Audience: Workshop Facilitators and Technical Leads

📋 Quick Reference¶

📅 Pre-POC Preparation¶

2 Weeks Before¶

1 Week Before¶

Day Before¶

🖥️ Environment Setup Checklist¶

Microsoft Fabric Overview¶

Source: Microsoft Fabric Lakehouse Overview

Azure Resources Required¶

Fabric Workspace Configuration¶

Workspace Name: casino-fabric-poc
Capacity: F64
Workloads:
  - Data Engineering: Enabled
  - Data Science: Enabled
  - Data Warehouse: Enabled
  - Real-Time Intelligence: Enabled
  - Power BI: Enabled

Lakehouses¶

📊 Data Generation Instructions¶

Generate All Data¶

cd data_generation

# Install dependencies
pip install -r requirements.txt

# Generate slot machine data (500K records, 30 days)
python generate.py slot_machine \
    --records 500000 \
    --days 30 \
    --output ../sample-data/slot_telemetry/

# Generate player profiles (10K records)
python generate.py player \
    --records 10000 \
    --output ../sample-data/player_profiles/

# Generate table games (100K records, 30 days)
python generate.py table_games \
    --records 100000 \
    --days 30 \
    --output ../sample-data/table_games/

# Generate financial transactions (50K records, 30 days)
python generate.py financial \
    --records 50000 \
    --days 30 \
    --output ../sample-data/financial/

# Generate security events (10K records, 30 days)
python generate.py security \
    --records 10000 \
    --days 30 \
    --output ../sample-data/security/

# Generate compliance data (5K records, 30 days)
python generate.py compliance \
    --records 5000 \
    --days 30 \
    --output ../sample-data/compliance/

Verify Data Quality¶

import pandas as pd
import os

# Check all files generated
data_dirs = ['slot_telemetry', 'player_profiles', 'table_games',
             'financial', 'security', 'compliance']

for dir_name in data_dirs:
    path = f'../sample-data/{dir_name}'
    if os.path.exists(path):
        files = os.listdir(path)
        print(f"{dir_name}: {len(files)} files")
        if files:
            df = pd.read_parquet(os.path.join(path, files[0]))
            print(f"  Sample rows: {len(df)}, columns: {df.columns.tolist()}")
    else:
        print(f"{dir_name}: MISSING")

📚 Session Facilitation Guide¶

🏗️ Day 1: Medallion Foundation¶

🌅 Morning Sessions (9:00 - 12:30)¶

💡 Key Teaching Points¶

❓ Common Participant Questions¶

Q: "Why not transform data immediately at ingestion?"

A: Bronze preserves the original data for auditing, debugging, and reprocessing. You can always rebuild Silver/Gold from Bronze. This is critical for compliance and troubleshooting.

Q: "How do we handle schema changes in source systems?"

A: Bronze uses schema-on-read with mergeSchema option. Discuss evolution strategies and how Delta Lake handles this automatically.

Q: "What about real-time data?"

A: Preview Day 2's Eventhouse content. Bronze can receive streaming data too, but Eventhouse is optimized for real-time queries.

👥 Hands-On Exercises¶

☀️ Afternoon Sessions (13:30 - 17:00)¶

💡 Key Teaching Points¶

👥 Hands-On Exercises¶

⚡ Day 2: Transformations & Real-Time¶

🌅 Morning Sessions (9:00 - 12:30)¶

💡 Key Teaching Points¶

❓ Common Participant Questions¶

Q: "When should we use SCD Type 1 vs Type 2?"

A: Type 1 for attributes that don't need history (e.g., email preference). Type 2 for business-critical attributes (e.g., loyalty tier, address for compliance).

Q: "How often should Gold tables refresh?"

A: Depends on use case. Daily for reports, more frequently for operational dashboards. Direct Lake enables near-real-time without scheduled refresh.

👥 Hands-On Exercises¶

☀️ Afternoon Sessions (13:30 - 17:00)¶

💡 Key Teaching Points¶

🎬 Demo Script - Streaming Producer¶

Run this in a separate terminal to simulate real-time data. Show participants how data flows through Eventstream to KQL.

import json
import time
import random
from datetime import datetime

while True:
    event = {
        "machine_id": f"SLOT-{random.randint(1000, 9999)}",
        "event_type": "GAME_PLAY",
        "event_timestamp": datetime.utcnow().isoformat(),
        "coin_in": round(random.uniform(1, 100), 2),
        "coin_out": round(random.uniform(0, 80), 2),
        "zone": random.choice(["North", "South", "East", "West"])
    }
    print(json.dumps(event))
    time.sleep(0.5)

👥 Hands-On Exercises¶

📊 Day 3: BI, Governance & Mirroring¶

🌅 Morning Sessions (9:00 - 12:30)¶

👥 Audience Expansion: Day 3 morning adds BI developers (2) to the architects (4).

💡 Key Teaching Points¶

❓ Common Participant Questions¶

Q: "What happens if Direct Lake falls back to DirectQuery?"

A: Discuss fallback triggers and how to avoid them (table size limits, unsupported DAX patterns, complex calculations).

Q: "Can we use Import mode instead?"

A: Yes, but you lose automatic refresh. Discuss trade-offs: Import is faster for complex DAX but requires scheduled refresh.

👥 Hands-On Exercises¶

☀️ Afternoon Sessions (13:30 - 17:00)¶

👥 Audience Expansion: Day 3 afternoon includes all teams (10+).

💡 Key Teaching Points¶

🎬 Demo Script - Mirroring¶

If SQL Server source is available:

1. Show CDC configuration
2. Create mirrored database in Fabric
3. Query mirrored data alongside Gold tables
4. Show replication lag metrics

If no source available:

1. Walk through UI steps with screenshots
2. Discuss architecture patterns
3. Show Microsoft documentation

👥 Hands-On Exercises¶

🔧 Troubleshooting Guide¶

Common Issues & Solutions¶

📋 Assessment Checkpoints¶

Day 1 End¶

Day 2 End¶

Day 3 End¶

📚 Participant Materials¶

Pre-Reading (Send 1 Week Before)¶

Reference Cards (Print/PDF)¶

Provide printed or PDF reference cards for:

• Delta Lake commands cheat sheet
• Common KQL queries
• DAX formula patterns
• Keyboard shortcuts for Fabric

Post-POC Resources¶

📝 Feedback Collection¶

Daily Feedback (End of Each Day)¶

Collect the following:

1. What worked well today?
2. What could be improved?
3. Pace rating (1-5): [ ] 1 [ ] 2 [ ] 3 [ ] 4 [ ] 5
4. Content clarity (1-5): [ ] 1 [ ] 2 [ ] 3 [ ] 4 [ ] 5
5. Any blockers for tomorrow?

Post-POC Survey (Send Within 1 Week)¶

📞 Contact Information¶

📋 Version History¶

📖 Related Documents¶