GraphConnect 2018

One of the best parts of my job is how much I get to learn about topics that I have previously only heard of. One of those things was graph databases and how they can be used and abused for connected data.


I was lucky enough to go to this year’s GraphConnect 2018 in New York City this September 20-21 at the Marriott in Times Square. As a person who grew up on Long Island and worked in New York City, this was an interesting homecoming for me.

The Conference

The conference was a well put together affair and I did learn a ton of stuff about how graphs are used in other industries, the intersection of graph databases, Machine Learning, Deep Learning and AI in general.


  • State of the Graph: Opening Keynote
  • The Present and Future of Artificial Intelligence and Machine Learning
  • Neo4j Graph Platform for Demographics Master Data at Citigroup
  • Predictive Analysis from Massive Knowledge Graphs on Neo4j
  • Neo4j Morpheus: Interweaving Table and Graph Data in Spark
  • Large-scale real time recommendations with Neo4j
  • DeepWalk – Turning Graphs into Features via Network Embeddings
  • How Graph Technology is Changing Artificial Intelligence and Machine Learning
  • Infinite Segmentation: Scalable Mutual Information Ranking on real world graphs
  • Machine Learning Algorithms in Neo4j


  • Neo4j for Data Science and Machine Learning
  • Graph Modeling Clinic

Taking a Walk

mountaineering copyI had a lot of fun at the conference, meeting new people and vendors selling everything from graph based analytics to ML training systems. The best talk of the whole conference was the “DeepWalk – Turning Graphs into Features via Network Embeddings“. This talk was a standing room only talk from one of the authors of the paper¬†and was a lot of fun going through¬†the paper and discussing how graphs can be used for feature extraction for our connected datasets and ML training grounds.

line_structure_relations_hierarchyThe first day was all talks while the next day was all learning. I took the Graph Modeling clinic to help me get over my propensity to design a data model like a RDBMS ERD diagram and the “Data Science and Machine Learning” course for the fun of learning more about Machine Learning and Model training.

All in all, a nice conference with interesting speakers and training without a lot of pushy sales people. I have to hand it to Neo4j, they put on a nice conference.