Synthetic Data Generator (Python)

This project generates synthetic data from scratch based on constraints and simple rules without requiring an original dataset.

I built this because in analytics, market research, and capstone projects, you often need realistic-looking data to prototype analysis, dashboards, or workflows, but can’t use real or proprietary data.

What this tool does

• Generates CSV datasets from a YAML specification
• Supports numeric ranges, categories, dates, and IDs
• Allows rule-based dependencies between fields
• Works for any topic (not tied to a specific domain)

The generator doesn’t assume anything about the data’s meaning it just follows the structure you define.

How it works (conceptually)

1. You describe the dataset shape in a YAML file (columns, bounds, rules)
2. The generator creates base values within those bounds
3. Rules override values where conditions are met
4. The result is written to a CSV file

Why YAML

YAML keeps the data definition readable and easy to change without editing Python code. Most changes happen in the spec, not the generator.

Limitations

• Rule-based dependencies (not statistical modeling)
• No guarantee of real-world distributions
• Optimized for clarity and flexibility, not scale

Example use cases

• Capstone projects
• Analytics prototyping
• Market research simulations
• Synthetic datasets for dashboards or demos

How to run

python synthgen.py