Spatiotemporal modeling tools for Python

This package provides tools for modeling and analyzing spatial and temporal autocorrelation in Python. It is based on the methods from the paper Spatial and temporal autocorrelation weave human brain networks. Included are methods to compute the following statistics:

• Compute TA-Δ₁ (i.e. first-order temporal autocorrelation)
• Compute SA-λ and SA-∞ (i.e. measurements of spatial autocorrelation)
• Lin's concordance
• Fingerprinting performance, from Finn et al (2015)

It will also generate surrogate timeseries for the following:

• Spatiotemporal model from Shinn et al (2022)
• Noiseless spatiotemporal model from Shinn et al (2022)
• Zalesky matching model from Zalesky et al (2012)
• Eigensurrogate model from Shinn et al (2022)
• Phase scramble null model

See complete documentation

Installation

To install:

pip install spatiotemporal

Otherwise, download the package and do:

python setup.py install --user

System requirements are:

• Numpy
• Scipy
• Pandas

Citation

If you use this package for a paper, please cite: Shinn et al (2022)

Contact

Please report bugs to https://github.com/mwshinn/spatiotemporal/issues. This includes any problems with the documentation. Pull Requests for bugs are greatly appreciated.

This package is actively maintained. However, it is feature complete, so no new features will not be added. This is intended to be a supplement for the paper, not a general purpose package for all aspects of spatiotemporal data analysis.

For all other questions or comments, contact m.shinn@ucl.ac.uk.