Try it in Python

Import NumPy, matplotlib, and scikit-learn: dataset, model, train/test split, and metrics.

Load the wine dataset and inspect its shape, classes, and feature names.

Split into training and test sets (25% test, fixed seed).

Fit `LogisticRegression` and print the learned coefficients for the first few features.

Predict on the test set, print accuracy, and plot the confusion matrix.