Knowledge graphs and graph databases are on everyone’s lips and are increasingly being used in companies. People who come into contact with the topic for the first time, inevitably think of visualizations of networks, in many cases of social networks. On the one hand, this is a good sign, since it underpins the thesis that semantic networks (in contrast to relational data models) are structured very much like humans think. On the other hand, it often gets in the way of further considerations as to what knowledge graphs should essentially serve.
Knowledge Graphs are Data
A Knowledge Graph is a model of a knowledge domain created by subject-matter experts with the help of intelligent machine learning algorithms. It provides a structure and common interface for all of your data and enables the creation of smart multilateral relations throughout your databases. Structured as an additional virtual data layer, the Knowledge Graph lies on top of your existing databases or data sets to link all your data together at scale – be it structured or unstructured.
‘Good’ and ‘bad’ graph visualizations
Visualizations can support the process of knowledge modeling, especially in the first phase of ontology creation, which is often characterized by collaborative and communicative processes. Even more obvious is the benefit of successful knowledge graph visualizations in the search, analytics, and discovery phases. How are things related, e.g. in which hierarchical relationship are they to each other? Graph visualizations can answer such questions directly. ‘Good’ visualizations serve a concrete purpose and can address specific issues or learning situations, e.g. in e-learning systems. ‘Bad’ visualizations show nothing but a network or a huge graph that leaves the user open-mouthed, who now has learned nothing more than that everything is very complex and highly interconnected.
Knowledge graphs are much more than ‘just’ visualizations
A closer look at the entire Linked Data Life cycle (see ‘Benefiting from Semantic AI along the Data Lifecycle‘) gives you a holistic view of the ‘knowledge graph’ topic. It soon turns out that visualization especially supports a specific phase of graph-based data management very well, namely the analysis of data. At the same time, however, it also becomes clear that knowledge graphs or the ‘semantic layer’ are much more than the representation of circles and arrows, which are intended to illustrate a semantic network of relationships:
- Knowledge graphs are at the core of any data virtualization strategy designed to support the highly scalable integration of heterogeneous data sources.
- At the same time, semantic knowledge graphs support broad initiatives to improve data quality and data standardization in enterprises.
- On this basis, projects based on machine learning methods can achieve significantly lower costs in the data preparation phase, even with smaller amounts of test data, as they can be enriched with consistent semantic metadata.
- And last but not least, knowledge graphs provide excellent support for conversational artificial intelligence, e.g. realize chatbots or smart helpdesks.
Conclusion
Visualizations such as PoolParty GraphViews or Ontology Explorer support some phases of the Linked Data Life Cycle, but they are not the ultimate purpose of a knowledge graph. Rather, companies with knowledge graphs can face fundamental data management challenges, such as: the integration and linking of unstructured and structured databases, or the success of conversational AI.
