Neo4j

What do you like best?

I've used neo4j for ontology-based KB generation task. So far, I have found neo4j graph-based data representation to be the best fit for such tasks. Apart from that, neo4j easy to learn and operate with.

What do you dislike?

Neo4j master-slave architecture, writes are always done on master. Overall write-performance issues relative to SQL databases. Garbage collection pauses.

Recommendations to others considering the product:

If you think your problem can be solved by graph based data, neo4j is worth a try

What problems are you solving with the product? What benefits have you realized?

I've used neo4j for the ontology-based KB generation task. The knowledge base consists of health care data with ontology representation as disease, symptoms, medication, organ/part of body associated etc along with their respective data and their relationship with each other. applying different graph-ml algorithm (like decision trees) on available data in graph to get a good accuracy of diagnosis. This also focuses on Knowledge Validation, Inferencing and Explanation/ Justification process of knowledge Engineering to maintain our healthcare knowledge base. Currently, we do not have CCDA's available with information of symptom of patients (chief complaint, the reason for visit, etc.). For the time being, we are collecting data from online sources and will use that to symptom-based disease diagnosis.Healthcare KB generation of clinical data using cTAKES NLP and Neo4j.

As name of the project suggests "all men must serve", demo code of the project is available publically in the link. So far, I have found neo4j graph based data representation to be the best fit for such tasks.