A NoSQL database has been designed to manage semi-structured and document-oriented data for the flexible storage and retrieval of historical character data and their relationships. This design enables quick and cost-effective adaptation to potential changes during historical research.
The database model consists of a main collection of documents called Nobility» Within this collection, there are documents called Person, which represent historical characters. The base structure of each Person document includes a list of 23 attributes and four arrays of subdocuments. As a semi-structured database, the base structure can be modified whenever necessary.
Each Person document can be linked to other documents of the same type through two attributes: idFather and idMother, which contain the unique identifier of, respectively, the historical character’s father and mother. It allows for the creation of a hierarchical genealogical structure for any character stored in the database.
Furthermore, each Person document is linked to four possible types of documents through the aforementioned arrays, providing both embedded and referenced document relationships, as commonly found in document-oriented databases (embedded documents and referenced documents).
The database is hosted in the cloud infrastructure of MongoDB (Atlas). Currently, remote access is provided to researchers’ devices. In the future, the database will be deployed within a public entity associated with scientific research and accessible through a web environment, available to researchers, developers, and the general public for data consultation.
This database represents an innovation in the field of historical research projects. Historically, most databases used in research projects have been relational databases. A study conducted within the Atlantocracies project surveyed the technical characteristics of databases used in 37 internationally renowned history-related projects. Over 70% of those databases were relational, while only 9% were NoSQL databases, all of which were graph-oriented. Therefore, the authors’ database model is presumably the first document-oriented model used for this type of research.
The main advantage and difference compared to other proposed solutions lie in its high flexibility. Historical research is a dynamic field, with new attributes, objectives, and questions emerging as information sources are explored. This leads to constant variations in the collected data for each object of study. Document-oriented databases allow for flexible data processing. In our specific case, although historical characters have a basic attribute structure, this structure can be modified, expanded, or reduced as needed. Thus, each stored historical character can have a completely different internal structure. In contrast, relational models require fixed table structures that all insertions must adhere to, and modifying them entails significant investments in time, effort, and money. Therefore, the presented model can quickly and easily adapt to any historical character-related information source. Similarly, the chosen database management system can efficiently handle the dynamic requirements of historical research.