Unified functional network and nonlinear time series analysis for complex systems science: The pyunicorn package

Area of Science:

• Complex systems analysis
• Data science
• Computational neuroscience
• Climatology

Background:

• Modern data analysis requires sophisticated tools for complex systems.
• Integrating complex network theory and nonlinear time series analysis is crucial for understanding large datasets.
• Existing software may lack comprehensive functionality for these integrated approaches.

Purpose of the Study:

• Introduce pyunicorn, an open-source Python package for unified complex network and recurrence analysis.
• Provide a versatile toolbox for constructing and analyzing functional networks from time series data.
• Facilitate the application of advanced network theory and time series analysis methods.

Main Methods:

• Object-oriented and parallelizable Python package.
• Construction of functional networks (e.g., climate networks, brain networks).
• Application of complex network measures (spatial networks, interacting networks, node-weighted statistics, surrogates).
• Nonlinear time series analysis using recurrence quantification analysis, recurrence networks, visibility graphs, and surrogate time series.

Main Results:

• Demonstrates the utility of pyunicorn through examples, primarily in climatology.
• Enables the investigation of statistical interrelationships in large time series datasets.
• Provides insights into nonlinear dynamics of complex systems from time series data.

Conclusions:

• pyunicorn is a powerful, flexible, and accessible open-source tool for researchers in various scientific domains.
• The package facilitates the combination of complex network theory and nonlinear time series analysis for novel insights.
• Its applications extend to climatology, neuroscience, and other fields dealing with complex time series data.