Spectral analysis in Python
The Python code that turns thousands of raw operando spectra into something a person, or a model, can reason about.
The problem
An operando experiment produces a long series of spectra, each tied to a moment in the battery’s cycle. Analyzing them by hand doesn’t scale, and every manual step is a place for inconsistency to creep in.
What I built
Python tooling for processing and analyzing spectral data, which also feeds the machine-learning work with Yuxin Chen’s group.
TODO(Sidney): the steps your code handles (e.g. loading instrument files, cosmic-ray removal, baseline correction, normalization, peak fitting, aligning spectra with electrochemistry), and the libraries you use.
Images
TODO(Sidney): a diagram of the pipeline stages. Use a schematic, not real data, until results are cleared.
What I learned
TODO(Sidney): what you'd tell someone starting to analyze spectra in Python.