It’s week two of four for my run publishing TWIN4J.
Last week in the UK we enjoyed Bonfire Night, a night that I have found makes absolutely no sense outside of the UK. On 5 November 1605, a man named Guy Fawkes was arrested while guarding explosives beneath the House of Lords in London. He was part of a conspiracy to assassinate King James I and his parliament.
Ever since, on November 5, bonfires and fireworks have been lit to celebrate the foiling of the plot. The tradition continues to this day. I wonder what we would have made of the events nowadays.
But hey, any excuse for fireworks and marshmallows by an open fire!
On to this week’s events in Neo4j, we’ve had another jam-packed week of content that I’m delighted to deliver to your inbox. We take a look at a huge announcement for the company, take tackle rugby data, social graphs, predicting mergers and acquisitions, and more.
Let’s dive in!
Featured Community Member: Grant Beasley
This week’s featured community member is Grant Beasley.
Grant Beasley – This Week’s Featured Community Member
Grant has been prolific in the last month, publishing a number of blog posts that feature Neo4j, Rugby Analytics, and Prog Rock. As someone in the intersection of the venn diagram of people who are interested in neo4j, music, and sports data, there was no way I could have selected anyone else for this week’s featured community member.
In a two part series on Prog Rock, Grant uses Neosemantics to import and explore 70’s Prog Rock data. Grant’s series on Graph Databases for Rugby Analytics uses python to import Opta Analytics data in XML format into Neo4j, before exploring the dataset using Cypher.
Grant is a Data Scientist at SaleCycle, where he utilizes big data to enhance client experience. SaleCycle help their customers increase conversion rates, understand their users, and drive loyalty.
Welcome, Patrick Pichette!
In the past week there has been some huge news for Neo4j as a company!
As Emil outlines in this post on the Neo4j Blog, we are delighted to welcome Patrick Pichette to the Neo4j Board of Directors as an independent board member.
Patrick is the former CFO of Google and serves as the Chairman of the Board of Twitter and Lightspeed. In addition, he serves as an independent investor, advisor, and board member to a number of startups and innovation networks around the world.
Graph Databases for Rugby Analytics
As I’ve already mentioned, Grant Beasley has published two posts in a series on exploring Rugby Analytics in a Graph Database.
In the post, Grant uses Python to import XML data into Neo4j, creating a graph of Fixtures, Teams, Players, and Referees. Each fixture has a number of events attached to it, which Grant uses to analyze carries, tackles, and more.
In Open Ownership’s third technology showcase, Gavin Chait of Whythawk demonstrates how to import the Open Ownership dataset into Neo4j.
Gavin has also published a Jupyter Notebook on Github in which he also sets up full-text indexes to query business entities.
Implementing Facebook Social Graph Using Spring and Neo4j
Amr Saleh has published a comprehensive article to Javarevisited on how to implement a Facebook style Social Graph using Spring and Neo4j.
This article builds on Amr’s previous post from June 2020 where he used Neo4j and Spring Data Neo4j to build an Instagram clone.
Amr starts by explaining the graph structure, before diving into the code and creating data objects, route handlers, and data repositories in Spring Data Neo4j.
Show Amr some love by throwing him some claps on his Javarevisited article. You can also watch a demo video for the project on Youtube.
Network Visualizations with SigmaJS and GRANDStack
The article teaches you to construct a Neo4j graph using data from OpenFlights, configure the Neo4j GraphQL library to serve the data, before finally demonstrating how to build a React application with Sigmajs to visualize the results.
The code is available on Github.
Neo4j Browser for Functional Visual Validation
Yoann Maingon over at Ganister PLM has written a blog post on how they use Neo4j Browser for Functional Visual Validation. The post details how Neo4j Browser allows Ganister PLM to visually make sense of their data. They take advantage of the forced graph layout to visually represent their data, and use features in Neo4j Browser to change node sizes, relationship line thickness, and colors.
Neo4j Browser to the Rescue!
Predicting Mergers and Acquisitions Using Graph-based Deep Learning
Over on arxiv.org, Keenan Venuti has published a paper on Predicting Mergers and Acquisitions using Graph-based Deep Learning.
The goal of Keenan’s work was to utilise GraphSAGE, a popular graph machine learning framework, to predict mergers and acquisitions (M&A) of enterprise companies. His results were promising. The ML model was able to correctly predict a merger or acquisition with an 81.79% accuracy on the validation dataset.
There is an implementation of GraphSAGE available in the Graph Data Science Libary.
Tweet of the Week
Finally, my favorite tweet this week comes from our very own Data Science Advocate Clair Sullivan.
As part of Clair’s weekly series “Bite-sized Neo4j for Data Scientists”, she has published video #14 showing you how to run community detection algorithms on Neo4j using the Louvain method.
It is Friday! That means it is time for the next installment of "Bite-Sized #Neo4j for Data Scientists!" Come check it out to learn how to do community detection in a graph using the Louvain method!https://t.co/xhdI257lyd#DataScience #MachineLearning— CJLovesData (@CJLovesData1) November 5, 2021
Clair publishes a new video every friday, so give her a follow if you like the video or are interested in more content like this!