# Olive oil EEM teaching artifacts

Source dataset: Venturini, Fluri & Baumgartner, *Dataset of Fluorescence EEM and UV Spectroscopy Data of Olive Oils during Ageing*, Mendeley Data, DOI: 10.17632/g6y69g8gwm.1, CC BY 4.0.

Generated files:
- `orange_olive_eem_features.csv`: Orange-ready feature table (728 rows).
- `olive_eem_features_with_pca.csv`: same table plus PC1/PC2 from Python.
- `olive_oil_eem_teaching_deck.html`: bilingual HTML slide deck.
- `olive_oil_eem_python_orange_teaching_video.mp4`: English-narrated, bilingual-on-slide teaching video.

Orange Data Mining demo workflow:
1. File → open `orange_olive_eem_features.csv`.
2. Select Columns → set `aging_class` as Target, keep `sample_id` and `source_file` as Meta.
3. Preprocess → Impute missing values + Normalize.
4. PCA → inspect ageing trajectory.
5. Test & Score → compare Logistic Regression, Random Forest, SVM.
6. Confusion Matrix → discuss which ageing classes are confused.

Caution for teaching:
- This CSV contains features derived from raw EEMs for demonstration.
- For a publication-quality workflow, add instrument correction, blank subtraction, scatter handling, validation split by oil identity, and independent test sets.