Modeling - Practical Neo4j - page 31

Figure 3-18. Using labels to represent location or address types
Figure 3-19 uses relationships to represent address types. Using this approach, new types of locations can also
be added to application design more easily. In addition, we can add properties to the relationship, such as “Greg's
Mailing Address”.
Figure 3-19. Using relationships to connect location or address types to a user
In addition to handling the model more elegantly, we could more easily connect other node types to these
locations if the scope of the application changes. Finally, we can use the Neo4j spatial plugin to handle geo searches
such as locations within a boundary.
Intent Graph
The intent graph seeks to map out a motivation or reasoning using a combination of other subgraphs from social,
consumption, interest, mobile and location. The intent graph might also be defined as the predictive graph in that
it uses those subgraphs to make predictions based on formulized intent. Based on those subgraphs, applications
can make suggestions or provide options that are in line with the calculated user intent and as such the value and
complexity of the intent graph is high.
For example, it would extremely valuable for Amazon—as well as other retailers—to understand how to ensure
adequate inventory and minimal time-to-delivery for any product they offer. While Amazon can factor in certain
events, such as popularity of a product, those factors provide a limited view as compared to coupling them with
connections, interest and location. To complete such a task with relational databases, the model would take a form
similar to the one shown in Figure 3-20 .