Iris SVM Classification Model -- 2

I have a clean, ready-to-use Iris dataset and need a complete Support Vector Machine classification pipeline built around it. The job is straightforward: tune an SVM (scikit-learn or similar) to separate the three Iris species, document the process, and hand over reproducible code and results.

Here’s what I expect:
• A concise notebook or script that loads the data, performs any minimal preprocessing you find beneficial, runs hyper-parameter optimisation, trains the final model, and outputs accuracy, precision, recall and the confusion matrix.
• A short write-up (markdown inside the notebook is fine) explaining why the chosen kernel and parameters work best, plus any insights you notice in the feature space.
• Saved model file so I can deploy or reload it quickly later.

I’ll provide the CSV as soon as we start, and I’m happy to test the notebook in my own environment, so please stick with widely supported libraries such as pandas, NumPy, scikit-learn and matplotlib/seaborn for visuals.

If you’ve built SVM classifiers before, this should be quick; quality, clarity and reproducibility matter more to me than flashy extras.