Graphs Are The Next Frontier In Data Science

GraphConnect 2018, Neo4j’s bi-annual conference, was held in New York City in mid-September. Read about what happened, and why graphs are the next big thing in data science.
By Reshama Shaikh, Data Scientist.
c
comments

GraphConnect 2018

 
GraphConnect 2018, Neo4j’s bi-annual conference, was held in New York City in mid-September. It took place right in mid-town Manhattan at the beautiful Marriott Marquis Times Square. I had the opportunity to attend some parts of the conference.

This was the schedule of events:

  • Keynotes and Sessions (Thursday)
  • Training Sessions (Friday)
  • Ecosystem Summit (Friday)
  • Community Hackathon (Saturday)

History of GraphConnect

 
The first GraphConnect conference was held in 2012. It is worth noting the community and user growth of Neo4j. I spoke to a participant who attended the first conference 6 years prior who shared that back then about 50 to 100 people attended. This year, over 1000 people attended.

About Neo4j

 
Neo4j efficiency is derived in using “pointers” to connect data, rather than “indices.” Traditional tabular datasets may contain “sparse matrices” if not all relationships between entities exist which slows down performance and is not optimally efficient.

Neo4j is open-sourced. There is a Community Edition that is available for free and also an Enterprise Edition which is available under a commercial license.

Conference

 
Neo4j founder Emil Eifrem and iconic data scientist Hilary Mason both delivered fantastic keynotes. Video recordings are publicly available for viewing all the keynote talks (about 90 minutes in duration).

Hello fellow Graphistas and aspiring Graphistas! Did you miss the #GraphConnect keynotes yesterday from @emileifrem and the amazing @hmason? Sit down and watch now!!

— Neo4j (@neo4j) September 21, 2018

Conference Highlights / Key Takeaways from Emil Eifrem

 
Seeing our Data as Networks

If our world is a network, we should look at our data as connected. #neo4j widely used in health insurance companies.

— Reshama Shaikh (@reshamas) September 20, 2018

NoSQL is Resource Intensive / Graphs are Efficient

Adobe has a creative cloud. Users can follow other creatives. Used to use noSQL. Was very resource intensive.
Realizing their data was increasingly connected, reduced # of servers with > results. #neo4j

— Reshama Shaikh (@reshamas) September 20, 2018

Popularity of Graphs

#graphconnect #neo4j popularity of graph databases pic.twitter.com/Heb1EuveEF

— Jaspreet Singh (@singhjaspreet) September 20, 2018

99% of the world’s airline ticket pricing runs through #neo4j, says our CEO @emileifrem. #graphconnect

— Lance Walter (@lancewalter) September 20, 2018

Healthcare Application

Healthcare Network Graph#neo4j pic.twitter.com/8n0lf1jzEA

— Reshama Shaikh (@reshamas) September 20, 2018

Hilary Mason

Packed room for ⁦@hmason⁩ kicking off #GraphConnect NYC! #neo4j pic.twitter.com/gPKCBC6obm

— Jesús Barrasa (@BarrasaDV) September 20, 2018

.@hmason says, "What does successful AI look like? It's boring. Like Google Maps, you don't have to know anything or think about the visualization of getting to your destination. That's what #AI success looks like."#graphconnect

— GraphConnect 2018 (@GraphConnect) September 20, 2018

Trainings

 
Thirteen different training sessions were offered as a separate conference event. Topics included modeling, development, data science and analysis for people at all levels of Neo4j experience:

New to Graphs

  • Neo4j Basics
  • New Features in Neo4j

Data Scientists and BI/Analysts

  • Graph Algorithms
  • Data Science and ML
  • Discovery and Visualization with Bloom

Architects, DBAs, and Data Modelers

  • Intro to Graph Modelling
  • Graph Modelling Clinic

Developers

  • Python web app development
  • Building apps on the GRANDstack
  • Modeling for Developers (refactoring, evolving, hands-on Cypher)
  • Cypher Tuning & Performance
  • APOC Extensions for Analytics and Operations
  • Graph-Based Natural Language Understanding

.@amyhodler and @JMHReif teaching centrality algorithms at #GraphConnect pic.twitter.com/zC4XluvNxV

— M. David Allen (@mdavidallen) September 21, 2018

If you ever have a chance to sit through a #neo4j Algorithms course with @amyhodler, do it! #GraphConnect

— Clair J. Sullivan, PhD (@cjIsALock) September 21, 2018

Ecosystem Summit

 
This was a private event to bring together the top Neo4j open source contributors, influencers and inspirational community leaders. Discussions were held in a smaller room. It was an intimate and casual discussion with Neo4j executives.

Lovveeeee these women. They inspire me!!! Thank you for coming and speaking at the Neo4j Ecosystem Summit!!!!@gabidavila (Developer Advocate, Google), @reshamas (Data Scientist and Organizer of WiMLDS, PyLadies NYC), @jumokedada (Founder, Tech Women Network) #Neo4j #GraphConnect pic.twitter.com/cpFfXxHsdt

— Karin Wolok (@askkerush) September 21, 2018

Balancing #opensource and strategizing revenue is challenging for all tech founders #GraphConnect @emileifrem pic.twitter.com/iGn4Mf85fH

— Reshama Shaikh (@reshamas) September 21, 2018

- The biggest competitors of #neo4j are all the people who do not know about this product.
- We need to reach users who can and absolutely need to use graphs for their data. @emileifrem #GraphConnect pic.twitter.com/VwG3mMTRLX

— Reshama Shaikh (@reshamas) September 21, 2018

Hackathon

 
On Saturday, September 22, there was a free community event, called Neo4j Buzzword Bingo Hackathon, which was a hackathon using the open-source software. Over 120 tech lovers attended this event which was hosted at Stack Overflow.

Graphs workshop in full swing at @StackOverflow #networks @wimlds @NYCPyLadies @neo4j pic.twitter.com/WGJkz3tf2X

— WiMLDS NYC (@WiMLDS_NYC) September 22, 2018

.@mkheck look at how awesome your daughter is!!! @JMHReif leading the 'Intro to #Neo4j' workshop at this year's #GraphConnect #GraphHack #hackathon at @StackOverflow pic.twitter.com/Rn2eVwInMa

— Karin Wolok (@askkerush) September 22, 2018

Social Media

 
Conference hashtags are:

  • #GraphConnect
  • #Neo4j

Twitter Handles

  • @neo4j
  • @graphconnect
  • @emileifrem (Founder)

Neo4j Company Milestones

  • 2000: Neo’s founders encountered performance problems with RDBMS and started building the first Neo4j prototype
  • 2002: Developed the first ever version of Neo4j
  • 2003: First 24×7 production Neo4j deployment
  • 2007: Formed a Swedish-based company behind Neo4j. Also open sourced the first graph database, Neo4j, under the GPL
  • 2009: Raised seed funding, $2.5M, from Sunstone and Conor and continued development
    • First Global 2000 Customer
  • 2010: Released Neo4j version 1.0
  • 2011: Raised A round and moved headquarters to Silicon Valley
  • 2012: Raised $11M Series B from Fidelity, Sunstone, and Conor
    • GraphConnect SF 2012GraphConnect, first conference on graph databases
  • 2015: Raised $20M Series C from Creandum with Dawn and existing investors
    • 2M+ Downloads of Neo4j
  • 2016: $36M Series D from Greenbridge Investment
  • 2017: Neo4j announces the Graph Platform, a connection-first approach to data query, visualization and analysis data.

Resources

  • Graph Databases (book by: Jim Webber, Ian Robinson and Emil Eifrem)
  • Neo4j Community
  • Neo4j Certification
  • Neo4j Users Slack team
  • Become a Neo4j Community Maven
  • Download Neo4j Desktop
  • The Neo4j Operations Manual v3.4
  • DB-Engines

Fun

 
Networking at a Networks Conference

People connecting at #GraphConnect. A year from now you might not remember bullet 3 on slide 15, but you might be running a project with someone you meet. Coffee is important. ;) pic.twitter.com/WS1oqfKmVk

— M. David Allen (@mdavidallen) September 20, 2018

“I drink in pubs that are older than your country.” ��
says a Brit to an American at #GraphConnect

— Reshama Shaikh (@reshamas) September 21, 2018

Fun in the DevZone

GraphConnect is getting weird and keynote hasn't even started .... Lol #GraphConnect2018 #Neo4j #DevZone #GraphConnect #GraphDatabases pic.twitter.com/BEatese2pL

— Karin Wolok (@askkerush) September 20, 2018

Learning Neo4j

.@jimwebber visualizing the “graph learning curve.” Seems unfamiliar and scary at first (because it’s different), then people get it and eventually love it. #neo4j #graphconnect pic.twitter.com/woMX1Ba0F0

— Lance Walter (@lancewalter) September 21, 2018

Original. Reposted with permission.

Related:

  • 7 Steps to Understanding NoSQL Databases
  • Modern Graph Query Language – GSQL
  • How To Learn Data Science If You’re Broke