Data Import
Data import turns files, reference entries, examples, or generated spectra into a SpectraSherpa dataset.
Main Sources
- File upload for CSV, JCAMP-DX, NumPy, MAT, and optional SpectroChemPy-backed vendor files.
- My Dataset for saved datasets in the current project.
- NIST Library for reference spectra.
- Synthesis for first-party synthetic FTIR datasets.
- Example datasets for starter workflows.
What to Verify
Every import should answer four questions:
- Which files were loaded?
- Which extension and reader path was used?
- What are the sample and variable counts?
- Which axis and units define the spectral dimension?
The Files, Metadata, and Data Matrix panels exist to make this visible before the dataset enters a workflow.
Calibration Caution
When spectra and reference values arrive from separate files, align by stable sample identity whenever possible. A row-order join can train a model that looks numerically successful while learning mismatched targets.