If you are dealing with data science, I bet you will find yourself doing a lot of prototyping. Prototypes often have very strict time and money limitations. The first lesson of prototyping is to approach every prototype as an MVP. The key idea behind MVP is to have just enough core features to show a working solution. Bells and whistles can be implemented later, as long as you are able to demonstrate the main idea behind your prototype.
Focusing on core features does not mean that your prototype should not have a pretty UI or stunning data visualizations. If those are the main strengths of your future product, by no means include them. To identify the core features of your product, you should think in terms of markets and processes.
Ask yourself the following questions to check whether a particular feature should be included in the MVP...
