Notebook ideas

[r-ford]

• The new "EOFs with NumPy" uses synthetic data to simplify things as much as possible, but it skips over some details when working with real data. Before getting to the notebook that uses the xeofs package and hides these details, I think it would be valuable to have a notebook that properly goes through the analysis on real data. Some things to highlight:

  • Weighting by grid cell area
  • Removing NaNs (and then reconstructing the grid after the analysis)

• Rotated EOFs: why, how, and an example of fixing a problem

• EOF/SVD analysis of coupled fields (probably following Bjornsson and Venegas 1997)