YAML
---
name: generating-test-data
description: Generates realistic placeholder data for development and testing. Use when you need sample datasets, mock API responses, or test fixtures.
---

# Quick Start

```python
from faker import Faker
import random

fake = Faker()

# Generate test users
users = [
    {
        "id": i,
        "name": fake.name(),
        "email": fake.email(),
        "created_at": fake.date_this_year().isoformat()
    }
    for i in range(1, 6)
]

Workflow

1. Identify data schema - Define fields, types, and relationships needed
2. Choose generation method - Faker library, manual patterns, or real data sampling
3. Generate base dataset - Create core records with realistic values
4. Add variations and edge cases - Include nulls, extremes, and special scenarios
5. Export in target format - CSV, JSON, SQL, XML as needed

Progress:

• Map required fields and data types
• Install generation tools (faker, mimesis)
• Create base data generation script
• Add edge cases and variants
• Export to required format

Examples

Example 1: Input: E-commerce product catalog (50 items) Output:

Python
products = [
    {
        "sku": f"PROD-{1000+i}",
        "name": fake.catch_phrase(),
        "price": round(random.uniform(9.99, 299.99), 2),
        "category": random.choice(["Electronics", "Clothing", "Home"]),
        "in_stock": random.choice([True, False])
    }
    for i in range(50)
]

Example 2: Input: CSV for user testing (100 rows) Output:

Python
import csv
with open('test_users.csv', 'w', newline='') as f:
    writer = csv.DictWriter(f, fieldnames=['id', 'name', 'age', 'city'])
    writer.writeheader()
    for i in range(100):
        writer.writerow({
            'id': i+1,
            'name': fake.name(),
            'age': random.randint(18, 80),
            'city': fake.city()
        })

Example 3: Input: API response with nested data Output:

Python
api_response = {
    "status": "success",
    "data": {
        "orders": [
            {
                "order_id": fake.uuid4(),
                "customer": fake.name(),
                "items": [
                    {"product": fake.word(), "qty": random.randint(1, 5)}
                    for _ in range(random.randint(1, 4))
                ],
                "total": round(random.uniform(25.00, 500.00), 2)
            }
            for _ in range(10)
        ]
    }
}

Best Practices

• Use Faker library for realistic personal data (names, addresses, emails)
• Use mimesis for performance-critical large datasets
• Seed random generators for reproducible test data: fake.seed_instance(42)
• Create data templates for common schemas (users, products, transactions)
• Include realistic distributions (80/20 rule, bell curves for ages/prices)
• Add intentional edge cases: empty strings, max lengths, special characters

Common Pitfalls

• Using obviously fake data like "Test User 1" that breaks realistic testing
• Forgetting to handle foreign key relationships in relational data
• Creating datasets too small to reveal performance issues
• Using production data patterns that leak sensitive information
• Not including sufficient data variety to catch edge case bugs