Maybe you know the following situation: you have a jupyter notebook for <whatever>. Then you want see the same analysis on a different dataset but don't want to lose your current results. So you make a copy of your notebook, and exchange the dataset. This might happen multiple times. Some time later, you have to change something in the code, and you do the same changes multiple times, since you have multiple coppies. Of course you could have moved the code from the notebook to a module and use the module in the notebook, but I believe that what I just described is not an uncommon scenario. This is what motivated me to write the library I want to promote here, which is called datasheet.