You hardly need python for models at this point. There are native julia implementations of linear regression, naive bayes, gradient boosting, random forest, PCA

You hardly need python for models at this point. There are native julia implementations of linear regression, naive bayes, gradient boosting, random forest, PCA, T-SNE and more. The python interop is just a stopgap