Decision Trees Notebook doesn't classify elements right

The notebook explaining decision trees, found under [here](https://github.com/JuliaPluto/featured/blob/main/src/machine-learning/decision-trees.jl) doesn't work properly:

- The tree partly wrong classifies some instances. Exp: with the default setting all features set to minimum value, the wine is classified as bad, although it should be good.
- At the bottom nodes, the tree contains two leafs with the same label (good), although all the instances that lead to a good wine, should be summarized together and the tree can stop at a level prior. (The reasons behind this are probably found in the implementation of the DecisionTree in the MLJ package and should be examined there)
- When the tree contains too many nodes, the text is unreadable anymore. some kind of wrap text or a similar approach should be implemented but CairoMakie doesn't offer this in graphs yet.