November 2014 Summaries

Medium, a social news and blogging platform, uses Neo4j to power its social functions. Medium chose Neo4j because it allows for the representation of users, posts, and collections as graph nodes with relationships between them, simplifying queries. Developer Evangelist Michael Hunger found that working with Neo4j's graph database felt natural, allowing him to easily store and retrieve data in a flexible manner. Graph Thinking is addictive, making it easy to identify connections between pieces of information, even in complex data. The story of how people came to appreciate the power of graphs has been shared by various community members, including Rik Van Bruggen, who found Neo Technology's vision resonated with him. A free copy of O'Reilly's Graph Databases ebook is available for download to learn more about graph databases and their applications.

LOAD CSV is a useful tool in Neo4j for importing datasets from various sources, including the internet. A live-coding session will demonstrate how to download a raw .csv file, clean it, visualize a data model, and write a Cypher query to import the data into Neo4j. This presentation aims to address common obstacles users face when dealing with real-world data in .csv format and provide best practices for using LOAD CSV. The session is designed to be followed along on one's own machine, with necessary tools and resources provided. Additional resources, including example Cypher queries and a Github repository, are available for further exploration.

Py2neo 2.0 has been released, marking a significant update after three years of development. The new version features a heavily refactored core, a cleaner API, improved performance, and enhanced support for Neo4j's labels, Cypher transactions, and other advanced features. The library now uses a more intuitive naming convention, with the `py2neo.neo4j` namespace deprecated in favor of `py2neo`. Additionally, new constructors for nodes and relationships have been introduced, allowing for greater flexibility and control over data synchronization between client and server. Py2neo 2.0 also integrates Cypher in a cleaner way, making it easier to execute queries and use transactions. The library has undergone various changes, including the introduction of `py2neo.legacy` for legacy code, new batch classes, and improved REST functionality.

The author shares their experience with using CoreOS as an operating system for running Docker containers, particularly with Neo4j as the database. They use a $5 Digital Ocean droplet to set up a CoreOS instance and install Neo4j Community Edition using a custom Docker image. The author highlights the ease of use of CoreOS and its built-in Docker features, which simplify the process of managing containers. Additionally, they demonstrate how systemd makes it easy to control and manage Docker containers, including restarting failed containers automatically. The author concludes by recommending CoreOS and Docker for their simplicity, scalability, and reliability, making them a great choice for server/cluster management and graph database applications.

We used KeyLines to visualize Neo4j's data, focusing on its GitHub repository, which has a vast amount of data flowing through it every second. The visualization highlights the relationships between individual contributors and files within a single repository, enabling a richer analysis of project data. By using a graph model with nodes representing contributors and files, and links representing commits, we can gain insights into team structures, expertise, and collaboration patterns over time. The application allows users to filter time ranges, view important files, and see who's been working on them, providing valuable information for teams and projects.

Graph databases can revolutionize the recruitment industry by providing a more efficient and effective way to store, manage, and analyze data about candidates and their relationships. Unlike traditional relational databases, graph databases are designed to show relationships and patterns between records, making it easier to identify key connections and potential matches for job openings. By leveraging graph databases, recruitment sites can quickly understand and classify candidates based on their background, skills, and interests, and proactively approach individuals who may be a good fit for certain roles. This can lead to better candidate matching, increased happiness among job seekers, and improved efficiency for recruiters. Graph databases also enable the use of machine learning and data analytics to identify patterns and trends in candidate data, providing valuable insights that can inform recruitment strategies and improve outcomes.

Switching from MongoDB to Neo4j — Nick Manning`: The author shares their experience of migrating their data from MongoDB to Neo4j, highlighting the benefits and challenges they encountered during this process. `Deep Dive on Fulltext Indexing with Neo4j — Stefan Armbruster`: This post delves into the world of full-text indexing in Neo4j, discussing its capabilities and best practices for implementation. `Using Graphs to Uncover Insider Trading Schemes — Linkurious`: The authors explore how graph databases can be used to identify patterns and connections that may indicate insider trading activities. `Connect Your Data Better with Neo4j — Rick Grehan`: This post provides an overview of the features and capabilities of Neo4j, demonstrating its potential for data integration and analysis. `Fraud Detection: Uncovering Connections with Graph Databases — Philip Rathle`: The author discusses how graph databases can be used to detect fraudulent patterns by analyzing connections between entities. `Flexible Neo4j Batch Import with Groovy — Michael Hunger`: This post provides a step-by-step guide on using Groovy to perform batch imports in Neo4j, highlighting its flexibility and efficiency. `Anti money laundering (AML): the network graph analytics approach — Scott Mongeau`: The author explores how network graph analytics can be used for anti-money laundering purposes by analyzing connections between entities. `How Graphs Revolutionize Identity and Access Management — Rik Van Bruggen`: This post discusses how graph databases can be used to improve identity and access management systems, enabling more efficient and secure authentication processes. `HR Analytics and Graphs: Job Recommendations — Linkurious`: The authors demonstrate how graph databases can be used for HR analytics, providing personalized job recommendations based on employee connections and behaviors. `Neo4j.rb 3.0! — Chriss Grigg, Brian Underwood, Andreas Ronge`: A new version of the Neo4j Ruby driver has been released, offering improved performance and features. `Release of Graphgen — Christophe Willemsen`: The author announces the release of a new tool for generating graph data, making it easier to create realistic graphs for testing and development purposes. `Beer Recommendations with User Based Collaborative Filtering — Michael Lam`: This post showcases how graph databases can be used to provide personalized beer recommendations based on user behavior and preferences. `Graphnote — Team Graphnote (Rails Rumble hackathon)`: A new project, Graphnote, has been developed during a Rails Rumble hackathon, demonstrating the potential of graph technologies for real-world applications. `Ebola Twitter Analysis — Swainjo`: The author analyzes Ebola-related tweets using graph databases, highlighting the effectiveness of this approach in understanding complex networks and patterns.

Mazerunner, an unmanaged extension for Neo4j, extends its capabilities to perform big data graph processing jobs while persisting the results back to Neo4j. It utilizes a message broker to distribute graph processing jobs to Apache Spark's GraphX module and persists the results in HDFS. Mazerunner enables scalable analysis of big data by providing a powerful tool for fast and efficient graph processing, allowing companies to gain competitive advantages with its ability to analyze large datasets quickly. The platform has shown promising results in running complex analyses on massive datasets, such as a Wikipedia dump, which resulted in a graph with over 10 million nodes and 104 million relationships, completing the analysis in under 3 hours on a laptop.