tssynth

Generate realistic synthetic time series data with customizable patterns, seasonality, and noise components. Perfect for testing time series algorithms, machine learning models, or generating sample datasets.

Features

✨ Flexible Generation: Create time series with various components:

• Linear, exponential, or custom trends
• Multiple seasonality patterns
• Configurable noise distributions
• Anomaly injection

🚀 Easy to Use: Simple, intuitive API built on numpy and pandas

import tssynth as ts

# Generate a time series with multiple components
data = ts.generate(
    length=365,                    # One year of daily data
    trend="exponential",           # Exponential growth
    seasonality=["weekly", "yearly"],
    noise_type="gaussian",
    noise_level=0.05
)

Installation

pip install tssynth

Documentation

For detailed usage examples and API reference, visit our documentation.

Quick Examples

# Generate a simple trend
basic_trend = ts.generate(length=100, trend="linear")

# Add seasonality and anomalies
complex_ts = ts.generate(
    length=1000,
    trend="exponential",
    seasonality="monthly",
    anomalies={"frequency": 0.01, "magnitude": 3.0}
)

# Export to pandas DataFrame
df = complex_ts.to_dataframe()

Development

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black .
isort .

License

MIT License - See LICENSE file for details.

Author

Will Judge (williamjudge94@gmail.com)