xmca 1.0.5.post1

Maximum Covariance Analysis in Python

Maximum Covariance Analysis in Python

Maximum Covariance Analysis (MCA) maximises the temporal covariance between two different data fields and is closely related to Principal Component Analysis (PCA) / Empirical Orthogonal Function (EOF) analysis, which maximises the variance within a single data field. MCA allows to extract the dominant co-varying patterns between two different data fields.

The module xmca works with numpy.ndarray and xarray.DataArray as input fields.

Testing

python -m unittest discover -v -s tests/

Core Features

• Standard MCA/PCA
• maximise covariance instead of correlation ==> Maximum Covariance Analysis (MCA)
• apply latitute correction to data fields to compensate for stretched areas in higher latitutes
• apply rotation of singular vectors
  • Orthogonal Varimax rotation
  • Oblique Promax rotation
• complexify data via Hilbert transform to inspect amplitude and phase information