A development team is exchanging data between a web application and a backend service. Each record can have a different set of fields, some fields contain nested arrays, and there is no fixed schema enforced. The team stores this data in files that use tags and key-value pairs to describe each element. How should this data best be classified?