Graph Model Proposals for Capturing Meta-information Within Professional Network Data
Călin Constantinov, Dorian Dogaru, Mihai Lucian Mocanu · 2020
Any successful analytics platform is typically backed-up by a fitting data persistence layer. In a world where storage cost is no longer an issue, deduplication and compact representations go against the idea of enabling the analysis of large streams of information. The focus should instead be on choosing a solution that naturally maintains the structure of data both when storing as well as when processing it. Thus, given their expressiveness and ease of use, graph databases have now become a widely-popular approach. In the context of professional network information, this paper presents a succession of graph data models that should enable faster retrievals while building an extensible foundation for enabling boundless improvements. By the means of a synthetically constructed dataset, several validations are performed, demonstrating how the querying process can be optimised for achieving improved execution times. While indeed expansive in terms of space consumption, the proposed techniques are highly-adaptable and are likely to benefit a large range of scenarios.