There is already `plaid.utils.stats`, but there is nothing to make a visual analysis of a `Dataset` We could propose : - scatter plot of each feature (sample id vs feature value) - pairplot of scalars - kdeplot