Gathers various notebooks and tools to assess generate synthetic data and test statistical diagnostics.
Data considered:
- time series (1D)
- spatiotemporal data (2D/3D)
Analysis:
- mean, variance
- autocorrelations
- spectral estimations
- confidence intervals are systematic
Leverages xarray, pandas objects and dask distributed.
Useful ressources:
- statsmodel: time series statistical analysis with pandas objects
Various interesting material:
- dask seeding stackoverflow post
- statsmodel ARMA generation
- what is a partial autocorrelation wiki link
- robust correlations with xarray and dask
- time series for scikit tools post
- scikit-learn
- dask-ml