MVPs are a core step in effective project development, and can be thought of as the smallest amount of work needed to meaningfully address your project's key question/purpose. That doesn't mean they necessarily answer the question, but instead that they provide something immediately functional and useful that can receive feedback and be built upon.
A helpful analogy: if you're designing a transportation vehicle, your final product might be a car, but your MVP on the way there is a skateboard. It's a useful tool that provides transport and is significantly easier to build, and your client can more quickly confirm that this is the type of product that they're looking for before you invest more time and effort into enhancing it.
In the context of Metis projects, MVPs will typically be simple results from a preliminary analysis or a functional baseline model/system that address the central project purpose and/or offer actionable takeaways. Their key prerequisite is access to data amenable to analysis/modeling, so they also serve as a check on your project progress. In its simplest terms, you can think of an MVP submission as "show us what you have so far that solves the problem".
- MVPs should be a short markdown file or jupyter notebook showcasing initial findings/models.
- Select a very small number of plots or relevant visuals (1 or 2 sounds perfect) from your project.
- Pull out the plots, give them concise captions/context, and write up your initial conclusions.
- Refer to the regression project MVP example for a handy, simple example.
Just show us something. Hopefully, you are looking backwards at things you have already done, picking something, sticking it up on GitHub, and then moving forward with whatever you are working on.