A knowledge graph allows you to construct a data model that simulates a real-world system and enter in data that lives within that model and can be accessed in a model-aware analytical process. Knowledge graphs build on the existing technology of graph databases, which provide mechanisms for persisting and analyzing graph and relationship-based data models like a knowledge graph. These data models are implemented in ArcGIS Pro and ArcGIS Enterprise through a product named ArcGIS Knowledge.
Entities in a knowledge graph represent real-world objects, concepts, or events such as a harbor, a test plan, or a repair. Relationships in the graph express how entities are associated with each other. Different transportation systems and the organizations and people who support them all meet at a harbor in different ways, for example. People, computers, and both physical and network resources are brought together to implement a test plan and assess the quality of the results. Customers, facilities, and supplies are all factors when repairing buildings, appliances, and transmission networks.
Building a representation of these real-world systems requires leveraging data that are primarily non-spatial: organizational structures, contractual obligations, material lifespan, payments, and documents. The knowledge graph you create allows you to discover how parts of the system are connected, which factors in the system have the biggest impact, and which hidden connections have more influence than expected.
Using ArcGIS Knowledge to store and analyze this data allows you to blend your library of tabular data and documents, as well as non-spatial analysis of the networks described by the knowledge graph, with spatial and temporal data and analysis. Consider the following scenarios, in which link analysis techniques such as assessing an entity’s centrality to the network can be combined with GIS analysis techniques including spatial statistics to provide a more complete view of the system: