Being able to see some stats on the resulting random forest can be useful to understand the dataset. Feature with most split, depth of trees...