Document Data Modeling: A Conceptual Perspective

Abstract

The growing availability of data and the increased popularity of NoSQL databases, that support the idea of managing unstructured or semi-structured data, motivate implementers to skip the phase of a conceptual view of data. However, document data stores belonging to the NoSQL group show a clear tendency of looking for some common feature among documents creating collections. This aspect motivates us to propose a model for the conceptual representation of a document data store based on UML class diagrams and mapping rules for its implementation. We also include a case study using Twitter data and show implementation using three data stores: MongoDB, CouchDB, and ArangoDB.