Final answer:
For storing unstructured data in a big data system, a key-value data store is most appropriate due to its flexibility and schema-less nature, which is ideal for handling various formats and sizes of unstructured data.
Step-by-step explanation:
When designing a big data system with poly-structured data that includes a significant amount of unstructured data, the most suitable type of data store to use is a key-value data store. Relational data stores are structured and typically not suitable for unstructured data, while generic lists of objects and XML data stores are more specific to certain data types and may not offer the scalability and flexibility needed for a big data system handling unstructured data. Key-value stores provide a highly flexible schema-less approach to storing unstructured data, which can vary in format, structure, and size.