Processing very large datasets using pyXem: STEM-DPC of ferromagnetic domains

This Jupyter Notebook was originally presented at the "ORNL/CNMS Virtual Workshop: AI for Atoms: How to Machine Learn STEM, December 7-10, 2020".

It shows how large scanning transmission electron microscopy - differential phase contrast (STEM-DPC) datasets acquired with a fast pixelated electron detector can be analyzed interactively using the open source python library pyXem.

A dataset is also included, with the "fe60al40_stripe_pattern.hspy" being the original one. Binned versions of this dataset is also included:

- Binned 2 x 2 in the probe dimension: "fe60al40_stripe_pattern_small_dataset.hspy"
- Binned 4 x 4 in the probe dimension: "fe60al40_stripe_pattern_very_small_dataset.hspy"

Python packages:

The notebook was run with

- pyxem 0.12.3
- hyperspy 1.6.1
- hyperspy-gui-ipywidgets 1.3.0
- jupyterlab 2.2.9
- ipympl 0.5.8

pyXem information: https://pyxem.github.io/pyxem-website/
HyperSpy information: https://hyperspy.org/

The data is from the journal article "Strain Anisotropy and Magnetic Domains in Embedded Nanomagnets": https://doi.org/10.1002/smll.201904738

The full open dataset and the data processing scripts: https://doi.org/10.5281/zenodo.3466590