What do you like best?
I started using Neo4J to represent various facilities ( of a startup that I work at) across the city as nodes and represent their relationships in terms of mode of commute available and the associated costs.
The brilliant thing about Neo4J, or graph DB for that matter of fact, is that it represents a graph like database as a graph and not a format manipulated to be used like a graph. This is turn makes it extremely practical for various graph based use cases like route optimization, connectivity networks etc
One more added advantage with using neo4j is the ability to use years of research done on graphs & its algorithms.
Now, what sets neo4J apart :
1) The cypher query language is very natural to write and easy to understand.
2) Allows many RDBMS like constraints to be applied to a graph.
3) The in-built query browser eliminates the use of 3rd party tools for simple graph visualizations
What do you dislike?
1) Not optimized for long traversal queries. Queries that require multiple routes inherently.
2) Very few DB drivers available, mostly are SDKs wrapped over the cypher query language (and communicates with the DB over HTTP & hence slightly slower).
3) cypher queries though simple to get started with, quickly turns into a complicated one due to very few methods.
Recommendations to others considering the product
Just one : It doesn't scale with multiple cores if you are using the community edition.
So before using it in a production environment make sure that the community edition would meet your scale, as the enterprise edition is pretty steeply priced.
What business problems are you solving with the product? What benefits have you realized?
We were trying to create a graph of various facilities and depots with weighted routes. What we essentially wanted to do was find the cheapest route from one facility to another ( via multiple hops)
Benefits of using neo4j are pretty obvious for us, we have created a graph and neo4j is a graph db, so there can not be a more natural fit for our usecase.