I am looking at the possibility of creating a graph database or multiple databases using neo4j. There are two or more very distinct purposes for these databases, and they are not related to each other except that a type of node common between them. Would it be better to use separate databases that just have a type of node in common with each other, or a single database where the node in question has multiple types of relations/edges connecting to nodes that otherwise have no relation nor any purpose in relating to each other? Also, each user will only have a single use case applicable to them, so the other use cases would not be relevant to them (the extra relation types and nodes will have no purpose for them). I am looking at this from multiple perspectives, including performance, storage space, memory usage, ease of creation, and maintenance. I am inclined to view a single database as being the best solution, partly based on past experience (I created a neo4j database for NASA that is part of one of their directorate's production environment now), but I could be wrong.
