Because Memgraph uses an in-memory storage engine and is implemented in C++ it's faster than Neo4j and more suited for analyzing real-time data, for example from a Kafka data stream. It also provides more support for implementing custom procedures in Python, Rust and C/C++ compared to Neo4j which is only doing Java.
That's a good point. Benchmarking different systems is always tricky to get right, but we'll get something out soon.
The focus of these benchmarks was to explain to our existing community how we improved things and share with the broader software engineering community has worked for us in the hope it could help others working on graph projects.
Dgraph is actually very different on many aspects like data model, query language, even overall focus & target use-cases.
In my opinion it will be tricky to have a fair comparison. You can have a look at the benchmarks they published against Neo4j. It seemed like there was a lot of disagreements on both side.
What other graph databases would you like us to have a look at?