How do you perform PCA on spectral data?

Question

I have spectral (FTIR) data for several molecules in .csv form, and I want to be able to visualize and classify these molecules using Principal Components Analysis (PCA) in Python.

There are 9 relevant .csv files (one for each molecule). In each .csv file, there are two columns: Wavenumber (inverse centimeters) and Transmittance (%). How can I take the data for all 9 molecules and then do visualization and classification using PCA? Again, in Python? Any links to tutorials or code sources that are able to do this for multiple signal data would be very helpful.

Thanks!

Answer 1

You may read and merge them into a single Pandas Dataframe.

With the wavenumbers in the columns and transmittances in 9 rows, apply PCA for dimensionality reduction and visualizations (example).

Alternatively, you may first extract features from the 9 spectra (max transmittance, min transmittance, centroid wavelength, etc.) and then apply the PCA in this feature space (example).