Graphs without semantics are not enough

For simple graph oriented data relationships, a non-semantic (or property graph) database approach might solve a single dimensional problem like: shortest path, one-to-many relationships, weighted elements, structured inter-relationships.

But rarely are problems and queries that simple. Real world data is highly complex, multi-dimensional and needs the powerful additional features of a semantic graph solution.
• SPARQL 1.1 query language support

• Linked Open Data support – with RDF

• Unstructured, free field text and document processing

• Geo-spatial queries

• Temporal queries

• Social Network Analysis

• Reasoning/Inferencing

• Machine learning

• Bayesian belief networks/probabilistic processing
• Prolog with complex rule and logic processing
• Federated, multi-database integration
• Predictive analytics
• Hadoop/Cassandra/MongoDB support for massive data/document stores
• Dis-ambiguation of terms, names, locations
• JSON/JSON-LD
• SOLR – native faceted search

Also, property graphs require the fixed definition of classes/objects, types and nodes – basically a fixed schema. If the data or data sources ever change, those changes need to be coded before any new data can be accessed.

Comparison of Property Graphs and Semantic Graphs