I'm quite comfortable using Python and additional ML/Visualization packages for ML and data mining. How do the free data mining software RapidMinder, KNIME, etc. compare to this approach, and do they have a practical advantage over using ML libraries? Are these big software packages more than fancy drag-and-drop which slows you down if you have programming skills?
What is the approach of people who do serious data mining and who possess programming skills to choose? Do they prefer to write code?