The folding is driven by thermodynamics, but in a sense, so is everything!

All of nature tends to move from higher energy to lower energy states. You can approximately calculate the energy of a protein structure, but you'll be wrong by enough that your error is bigger than the difference between candidate structures. To calculate the energy with sufficient accuracy, you need to use quantum mechanics using, e.g., density functional theory.

> -they just consider the parts of the chain as having well-defined positions in 3d-space, and add up energy from pairwise interactions between the parts that end up close to each others.

Yes, but also if you want an accurate calculation of that energy, you need to use quantum mechanics. The nuclei of the atoms have definite positions, but electrons do not

Tableau is really good at linking to an underlying data source like a cloud database and creating visualizations that update automatically as you get new data. It has some support for analysis but not much, so for complicated stuff you want to preprocess it somewhere else