Semi-structured data is a category of data organization that does not conform to the formal structure of tabular data models associated with relational databases, yet contains tags or markers that impose a logical hierarchy and provide semantic meaning to its elements. As an academic concept, it concerns the study of data models, query languages, and management systems designed to handle data formats like XML, JSON, and other self-describing structures prevalent in web data and NoSQL databases, representing a spectrum between unstructured text and rigidly structured data.