Restaurant POS Data Analysis Project

Overview

This project demonstrates advanced data cleaning, processing, and visualization techniques using restaurant Point of Sale (POS) data. Using Python, SQL, and modern AI tools including Claude 3.5 Sonnet, I analyzed operational patterns to derive actionable business insights for restaurant management.

Data Cleaning Process

The project began with raw POS data containing common real-world issues such as missing values, price inconsistencies, and duplicate transactions. Through systematic cleaning and validation, I transformed this raw data into a reliable dataset for analysis.

Before Cleaning:

After Cleaning:

Key cleaning steps included:

• Standardizing price formats and correcting anomalies
• Handling missing server IDs, payment types, and table numbers
• Removing duplicate transactions
• Converting data to SQL format for efficient querying

Key Visualizations and Insights

1. Order Volume by Location and Hour

Key Findings:

• Peak hours show 3x higher order volume than off-peak periods
• Distinct patterns between lunch (1 PM) and dinner (5-6 PM) peaks
• Location-specific trends inform staffing needs

2. Daily Order Patterns

Key Findings:

• Weekend order volumes average 33% higher than weekdays
• Consistent afternoon lull across all locations at 3 PM
• Clear peak periods during lunch (1 PM) and dinner (5-6 PM)

3. Average Ticket Analysis

This visualization demonstrates the importance of working with real data for certain analyses, as synthetic data may not capture true variations in customer behavior patterns.

Business Impact

The analysis revealed several actionable insights for restaurant operations:

• Optimize staffing by adjusting worker hours to match 33% higher weekend volumes
• Align staffing with peak hours (3x normal volume) to balance service quality and labor costs
• Adjust location-specific staffing based on unique lunch and dinner rush patterns

Tools & Technologies Used

• Python (Pandas, Matplotlib, Seaborn)
• SQL
• Claude 3.5 Sonnet for AI-augmented analysis
• Data Visualization Libraries

For detailed analysis methodology and findings, see ANALYSIS.md.

Project Structure

restaurant-analysis/
├── .gitignore
├── .env
├── README.md
├── ANALYSIS.md
├── LICENSE
├── requirements.txt
├── images/
│   ├── POS_CSV_pre-cleaning.png
│   ├── POS_SQL_post_cleaning.png
│   ├── order_per_hour_heatmap.png
│   ├── daily_hourly_patterns.png
│   └── daily_average_tickets.png
└── src/
    ├── utils.py
    ├── generate_pos_data.py
    ├── ave_ticket_by_day_2.py
    ├── heatmap_2.py
    └── each_day_hourly.py

Contact

For collaboration opportunities, reach out via symmetry1@live.com

© 2025 Heath Hoppus

License

This project is licensed under the MIT License - see the LICENSE file for details.