When embarking on a machine learning project that involves exploratory data analysis and model prototyping, which cluster type would best suit your needs?