Oh Leonhard, Where art thou?
"Most NoSQL stores focus on throughput and resilience to failure. The trade-off in achieving those goals has been a substantial decrease in the sophistication of data models compared to RDBMS systems, and a corresponding increase in processing frameworks to deal with large volumes of unconnected data. But (sometimes) persistent hashmaps and map-reduce clusters aren't the only option for dealing with interesting data and exploiting emergent intelligence: graph theory provides better approaches.
In this talk, I'm going to tell you about graph databases, the most esoteric and powerful quadrant of the NOSQL universe. I'm going to show how datastores like Neo4j can help us as developers to gain tremendous insight into our transactional data, and make a case for graphs to be part of your toolkit by showing just how fast graph processing is. There'll also be some terrible jokes about how Euler's model is actually 200 years more mature than the enterprise-y relational model, which you'll be able to use to embarrass your boss into moving your data into the 21st (or back to the 18th) century."
