Low quality Data Science code is not a fault of Jupyter. The Jupyter allow you to load big chunk of data or some large model only once, and then use it for exp

Low quality Data Science code is not a fault of Jupyter. The Jupyter allow you to load big chunk of data or some large model only once, and then use it for experiments in other cells. It is hard to replace this feature with plain `*.py` file. For me, this is the killer feature.