Sample Data for Microsoft Fabric Casino/Gaming POC¶

This directory contains pre-generated sample datasets for demonstrating and exploring the Microsoft Fabric Casino/Gaming POC without needing to run the full data generation pipeline.

Quick Start¶

The sample data is ready to use immediately. Simply load the CSV files from the bronze/ directory into your preferred tool (Excel, Power BI, Fabric notebooks, etc.).

# Example: Load in Python
import pandas as pd

slot_data = pd.read_csv("bronze/slot_telemetry_sample.csv")
player_data = pd.read_csv("bronze/player_profile_sample.csv")

Directory Structure¶

sample-data/
├── README.md                    # This file
├── generate_samples.py          # Script to regenerate sample data
├── bronze/                      # Pre-generated sample CSV/Parquet files
│   ├── slot_telemetry_sample.csv
│   ├── player_profile_sample.csv
│   ├── table_games_sample.csv
│   ├── financial_transactions_sample.csv
│   ├── security_events_sample.csv
│   └── compliance_filings_sample.csv
└── schemas/                     # JSON schema definitions (symlink)

Sample Datasets¶

1. Slot Machine Telemetry (slot_telemetry_sample.csv)¶

Records: 90 rows (pre-generated) / 500 rows (with generator) Description: Slot machine events following SAS protocol patterns

Key Features for Demo: - Includes 5+ jackpot events for payout analysis - Mix of carded and uncarded play - Various denominations and machine types - Error/tilt events for maintenance analysis

2. Player Profiles (player_profile_sample.csv)¶

Records: 25 rows (pre-generated) / 100 rows (with generator) Description: Loyalty program member profiles with masked PII

Key Features for Demo: - Realistic tier distribution (40% Bronze, 30% Silver, etc.) - VIP and Diamond tier players for high-value analysis - PII properly masked for compliance demonstration - Self-exclusion flags for responsible gaming

3. Table Games (table_games_sample.csv)¶

Records: 40 rows (pre-generated) / 200 rows (with generator) Description: Table game transactions and player ratings

Key Features for Demo: - Multiple game types with realistic distributions - Buy-in and cash-out transactions - Player rating data for host assignments - Marker/credit transactions

4. Financial Transactions (financial_transactions_sample.csv)¶

Records: 30 rows (pre-generated) / 200 rows (with generator) Description: Cage and cashier transactions

Key Features for Demo: - 10+ transactions above CTR threshold ($10,000) - Suspicious activity flags for AML review - Wire transfers and marker transactions - Check cashing with ID verification

5. Security Events (security_events_sample.csv)¶

Records: 30 rows (pre-generated) / 150 rows (with generator) Description: Security and surveillance event log

Key Features for Demo: - Access control events (granted/denied) - Camera alerts and surveillance events - Incident reports with categories - Exclusion check events

6. Compliance Filings (compliance_filings_sample.csv)¶

Records: 20 rows (pre-generated) / 50 rows (with generator) Description: Regulatory compliance filings (CTR, SAR, W-2G)

Key Features for Demo: - Mix of CTR, SAR, and W-2G filings - CTR filings above $10,000 threshold - SAR filings with structuring patterns - W-2G jackpot tax reporting

Data Characteristics¶

Referential Integrity¶

All sample datasets use consistent player_id values, allowing you to join across datasets:

-- Example: Join player profiles with slot activity
SELECT
    p.loyalty_tier,
    p.player_id,
    SUM(s.coin_in) as total_wagered,
    COUNT(DISTINCT s.session_id) as sessions
FROM player_profile_sample p
JOIN slot_telemetry_sample s ON p.player_id = s.player_id
GROUP BY p.loyalty_tier, p.player_id

Date Range¶

All sample data is generated for January 2024 (2024-01-01 to 2024-01-31).

Reproducibility¶

Sample data was generated with seed 42 for reproducibility. Running generate_samples.py with the same parameters will produce identical output.

Edge Cases Included¶

The sample data intentionally includes edge cases for testing:

• Jackpots: At least 5 jackpot events in slot data
• CTR Threshold: 10+ transactions >= $10,000
• Suspicious Activity: Flagged transactions for SAR review
• High-Value Players: Diamond/Platinum tier members
• Security Incidents: Critical and high-severity events
• Self-Exclusion: Players with responsible gaming flags

Regenerating Sample Data¶

To regenerate the sample data with different parameters:

# Install dependencies (if not already installed)
pip install -r ../requirements.txt

# Generate with defaults (seed=42, CSV format)
python generate_samples.py

# Generate with custom seed
python generate_samples.py --seed 123

# Generate larger datasets
python generate_samples.py --slots 2000 --players 500

# Generate both CSV and Parquet
python generate_samples.py --format both

# Generate to custom location
python generate_samples.py --output ./my-data

# Full customization
python generate_samples.py \
    --seed 42 \
    --slots 1000 \
    --players 200 \
    --tables 400 \
    --financial 400 \
    --security 300 \
    --compliance 100 \
    --start-date 2024-06-01 \
    --end-date 2024-06-30 \
    --format both

Command Line Options¶

Using with Microsoft Fabric¶

Upload to Lakehouse¶

1. Open your Fabric workspace
2. Navigate to your Lakehouse
3. Click "Get data" > "Upload files"
4. Select the CSV files from bronze/
5. Files will be available as Delta tables

Load in Notebook¶

# Fabric Spark notebook
from pyspark.sql import SparkSession

# Read uploaded files
slot_df = spark.read.format("csv")\
    .option("header", "true")\
    .option("inferSchema", "true")\
    .load("Files/sample-data/bronze/slot_telemetry_sample.csv")

# Or if converted to Delta
slot_df = spark.read.format("delta").load("Tables/slot_telemetry")

Power BI Connection¶

1. Open Power BI Desktop
2. Get Data > Text/CSV
3. Navigate to CSV files
4. Transform and load as needed

File Sizes¶

The pre-generated samples are smaller representative datasets for quick demos:

For larger datasets, use the generator script:

# Generate full-size samples (recommended for POC demos)
python generate_samples.py --slots 500 --players 100 --tables 200 \
    --financial 200 --security 150 --compliance 50 --format both

Schema Reference¶

JSON schema definitions are available in the ../data_generation/schemas/ directory:

• slot_telemetry_schema.json
• player_profile_schema.json
• table_games_schema.json
• financial_transaction_schema.json
• security_events_schema.json
• compliance_filing_schema.json

These schemas can be used for data validation in Fabric pipelines or Great Expectations.

Questions?¶

For issues with sample data generation, check:

1. Python dependencies are installed (pip install -r requirements.txt)
2. Running from the correct directory
3. Sufficient disk space for output

For POC-related questions, see the main README or CONTRIBUTING guide.