Designing a schema will vary according to the scenario of data and operations that are being used. So, once you have designed and implemented your graph model, you need to use the appropriate schema to retrieve interesting information and patterns from the graph. The tool of preference would be Cypher, as it is essentially built around graphs. Let's look at some design scenarios.
The data entities can have different levels of classifications, for example, different groups contain a given user, with different roles for each group and a user being part of several groups. The user can take up various roles in multiple groups apart from the one they are a member of. The association that exists between the user, their groups, and the roles can be depicted with the help of a hyper edge. This can be implemented in a property graph model with the help of a node that captures an n-ary relationship.