eSynergy Solution’s Community Manager, Adele Green talks with Jim Webber, Chief Scientist at Neo Technology about graph databases, community and the future of Neo4j.
Tell us about yourself?
I’m Neo4j’s Chief Scientist. Previously I was a Director at ThoughtWorks where I worked a lot on RESTful systems and large-scale distributed systems. In the old days I worked with marvellous Geordie startup Arjuna Technologies just after I was sprung from a wonderful life as an academic.
Why Graph databases?
For two reasons. Firstly they are powerful. The first time I happened to bump into graphs (an early version of Neo4j in fact), it allowed me to solve a problem (upsell and recommendations for Telco) in a matter of hours when we thought it would have taken months with traditional tech. Secondly, it really irritated the old skool’ types who had planned to take months to deliver that functionality. What’s not to like?
Why did you join Neo4j?
That was a tough call. Prior to Neo4j I was working with ThoughtWorks, a rather splendid IT consultancy and I felt pretty much at home there.
But I had started to see the potential for graphs and had used Neo4j in several substantial PoCs. I’d also been a contributor to Neo4j during that time. The Neo4j founders raised a bit of capital, and asked me if I’d like to switch to full time contributions. Oh, and with a massive pay cut :-/ I accepted, and 4.5 years later the rest is history.
What makes Neo4j’s offering to the market so special?
From a business point of view, it’s the first, most mature only enterprise-grade (ACID transactional) graph database out there now. From a technical point of view, the Cypher query language opens up graph theory to a wide range of developers and analysts. That’s truly powerful. SQL never lived up to its promise in that respect, and the other NoSQL languages – despite being competent – really only give you key-value interactions. But by opening up graphs through the humane Cypher language something extraordinary has happened: a stampede of people wanting to make sense of their connected data.
Could you give us the best example you have seen of Neo4j benefit a business?
Anywhere that you have connected data, which is pretty much everywhere. For example you could use Neo4j to provide an MDM-style meta-index over your existing systems, bringing all the disparate data together under one umbrella. Or you could tackle something domain specific: fraud detection, recommendations, routing, and their equivalent algorithms in many domains are easily expressed and exploited as graph problems.
When would you not use Neo4j?
Graph databases depend on structure. While that structure can be as uniform or irregular as you like, structure is what makes them powerful and fast. In a totally unstructured domain (of which there are very few), graphs aren’t all that useful. But once you have learned, inferred, or extracted structure from that domain then it’s graphs all the way.
What’s the best way for an organisation or individuals to get involved?
Neo4j has an enormous community worldwide. Here in London where I’m based our meetup group is in excess of 2,000 people. We meet several times per month. There are two sessions at the Neo4j office in Bankside where we do smaller hands-on workshops (e.g. language specific things or domain specific things) and we do a large meetup on the last Wednesday of every month at SkillsMatter where our community members tell their stories.
If you’re local that’s an easy way to get stuck in. We also buy the beers afterwards 🙂
There are lots of these meetups around the world, there’s likely to be one near you. If not, then we have a very active community on StackOverflow and a lively and friendly Google group too.
What’s the best piece of advice you can give a company who are looking to utilise a graph database in order to make their project successful?
Talk to us! Graph tech is still relatively new, and we’ve a lot of experience to share. Don’t wait until you’re confused or stuck, talk to us early on and let us help you. Neo4j is awesome stuff and you can be very successful with it, but learn from our mistakes, don’t make your own.
Oh, and read the book that Ian Robinson and I wrote with Emil Eifrem. It’s called Graph Databases, published by O’Reilly. You can get a free, full ebook version athttp://graphdatabases.com
Where do see the future of graph databases and Neo4j?
We’re going big. The analysts are saying that within two years a quarter of large enterprises will have deployed Graph Databases, and since Neo4j is 80+% of that market it means a lot of people will be using Neo4j. The category itself is expanding enormously both in Neo4j’s space (graph storage and query) and in the graph processing space (think: Hadoop for graphs). This is a massive growth area for data in the coming decade.
As a science-bod in Neo4j, this means we’re going to be doing some amazing and challenging R&D work to keep up with the ambition of our users!
Keywords: Jim Webber