April 2020 Summaries

Neo4j Innovation Labs is a methodology used by Neo4j to help companies accelerate innovation and transformation through graph thinking. The labs are collaborative exercises that encourage customers to think about processes in a connected way, rather than siloed and stale. The process involves generating use cases, data modeling, prototyping, and creating something tangible. This approach helps companies to achieve more in three and a half days using the Neo4j methodology than they have done on their own in a year. By adopting graph thinking, companies can overcome traditional limitations and tap into new possibilities for innovation and transformation. The Innovation Lab sprint is a three and a half day workshop that uses data science and graph technology to propel a way of thinking about processes. It provides a unique approach to solving complex problems by exploring connected data narratives.

The Neo4j community has been actively sharing knowledge and projects related to graph databases, with a focus on applications such as contact tracing during the COVID-19 pandemic. Various members have showcased their work, including de-duplicating ingredients in the BBC goodfood dataset using Cypher queries, building COVID-19 contact tracing graphs, exploring virus taxonomy, and demonstrating how stored procedures can be used in Neo4j 4.0. Additionally, there are updates on GORM for Neo4j, a Python wrapper to call procedures from the Graph Data Science Library, and a paper on using Neo4j for Open Linked Cultural Data. The community has also recognized Thomas Deely as a featured member, who organized the Knowledge Graph Conference and brought together experts in various fields to work on projects related to the United Nations Sustainable Development Goals.

The text discusses the challenges of sharing internal information and knowledge in a distributed team environment, where traditional methods like Slack and Microsoft Teams can make it difficult to find relevant data. A solution is proposed by using Neo4j to build an enterprise knowledge graph, which allows for the identification of relationships between entities such as people, skills, and files, making it easier to analyze and gain insights from the data. The authors demonstrate how this approach can be applied to a real-world scenario, where they built a Slack knowledge graph to determine user skill levels dynamically by analyzing their interactions. The architecture is designed to handle large amounts of data from multiple sources, including Slack, Github, Google Drive, and Salesforce, and provides a custom dashboard for users to explore the data.

Spring Data Neo4j RX is a full-fledged Spring application that leverages Spring Data's ecosystem, providing mapping and conversion support built-in. The library brings all mapping capabilities within the framework, making it easy to work with an instance of Neo4j created for use within a Spring Boot application. It supports immutable entities and reactive programming model, allowing users to take control over the stream of data they're querying for. The library bases its connections on the Java Driver, ensuring stable communication and utilizing the driver's type system reflecting the database's one. Users can access their domain by using a repository for the aggregate root as described in Domain Driven Design, or have more control over the querying and mapping process with the Neo4jTemplate and Neo4jClient. The library also includes CypherDSL, an experimental API that provides a public surface and features of the DSL, allowing users to try it out despite potential changes. To get started, users can create a new Spring Boot application using start.spring.io and add the Spring Data Neo4j RX starter dependency, then model their domain according to the documentation and create a suitable repository definition to access their data.

Graphs are being used in various ways to combat COVID-19, including tracing connections between people and identifying clusters of elevated activity, as well as drug research and repurposing. Graph databases like Neo4j are providing a platform for researchers to instantly access integrated knowledge from coronavirus data, enabling them to identify potential treatments and track the spread of the virus more effectively. The company is also offering free resources, licenses, and support to individuals and startups working on COVID-19 projects, as well as promoting visibility of these projects through its Graphs4Good platform. Individuals can get involved by joining existing projects or setting up their own, bringing technological proficiency, domain knowledge, or caring for those around them. The company is encouraging collaboration and offering various ways to help, including online training courses and sandboxes where people can experiment with graph databases in a safe environment.

This week's community news includes the release of Spring Data Neo4j RX 1.0, which utilizes a new reactive driver architecture in Neo4j and features such as improved CypherDSL, optimistic locking, and support for primary and secondary labels. Gerrit Meier, this week's featured community member, launched the new library alongside Michael Simons. Joe Depeau discussed how graph algorithms are used in banking, including using PageRank, Weakly Connected Components, and Node Similarity algorithms to explore a fraud graph. The Graphs4Good GraphHack COVID-19 edition has concluded, with projects such as Exposure Tracker, Project Domino, and Epidemic Simulator showcasing the potential of graph analytics for social good. Inzamam ul Haque shared tips for optimizing Cypher queries, while Max De Marzi built a Multi-Level-Marketing program using graphs, and Karin Wolok ran over the submitted projects.

Graph data science is increasingly being used by businesses to combat the growing problem of fraudulent schemes, despite sophisticated fraud detection tools. Traditional models that flag high-risk identities often miss synthetic identity theft and fraud rings, which can result in significant financial losses. Graph data science helps turn this pattern around by augmenting existing analytics and machine learning pipelines, increasing the accuracy and viability of existing fraud detection methods. By analyzing network structures using searches, queries, and graph algorithms, graph data science improves the accuracy of fraud predictions, enabling financial services firms to detect more fraud in their existing data without changing their ML pipeline. This approach also unlocks new insights into patterns indicative of fraud, allowing for real-time operational fraud detection systems to be incorporated and top-line benefits achieved.

This week's video featured Greg Woods teaching about the new reactive Neo4j drivers, while Alicia Frame officially launched the Graph Data Science Library. Rik Van Bruggen continued his series on supply chain management with graphs, and Lucas Moda created a graph of São Paulo's subway system using Cypher. The community also celebrated Global Graph Celebration Day, and featured Steven Mogg as a highlighted member for his contributions to the Neo4j community. Additionally, Michael Simons shared examples of using Spring Data Neo4j RX with JDK 14, while William Lyon wrote a deep dive into GraphQL resolvers used in GRANDstack, and Adam Cowley published updates to vue-neo4j. Markus Günther appreciated a Twitch live stream by Michael Simons, which gave an introduction to Neo4j.

The latest scalability features in Neo4j 4.0 include unlimited vertical and horizontal scalability, database sharding, and federation, which can be demonstrated to colleagues or bosses by confidently answering the question "But, does it scale?" with a simple "Yes, it does." To protect one's reputation as an expert, it is recommended to subscribe to the Neo4j YouTube channel for regular updates on graph technology.

Graph technology is being adopted by enterprises to gain a competitive advantage and drive insightful decisions, particularly in areas such as decision support with knowledge graphs, efficiency of processing with graph accelerated ML, accuracy with connected feature extraction, and transparency into AI decision-making. These technologies are crucial for valuable AI applications, especially in critical business areas like network security, where preventing attacks and securing corporate networks is a top priority. Companies like Fujitsu Laboratories and Larus, a Neo4j partner, are developing new graph AI technologies, such as Deep Tensor, which can automatically extract relevant features from graph data and provide additional explanations for AI-generated findings. By leveraging these technologies, organizations can enhance the efficiency of their business operations by extracting actionable insights from connected data, ultimately driving business advantage through informed decision-making.

The Neo4j community has been actively engaging in various projects and sharing their knowledge through videos, blog posts, and social media. Michal Bachman used the Hume insights engine for contact tracing and smart quarantine during the coronavirus pandemic, while Rik Van Bruggen continued his series on supply chain management with graphs. TasoX demonstrated a tool for importing Windows Events into Neo4j, and NEuler was updated to use the Graph Data Science Library. Adam Cowley introduced new features in the latest release of Neo4j Desktop, including a first graph experience for new users and file functionality. The community also highlighted the work of Elena Kohlwey, a Certified Neo4j Developer and Neo4j Ninja, who has been actively contributing to the community and helping others through ask-me-anything sessions and feedback on the Ninja Program. Additionally, there were updates on Windows Events & Sysmon Visualization using Neo4j, Graphlytic for Fraud Detection, and BloodHound 3.0 on Neo4j 4.0.

Bloom is a graph visualization, exploration and communication tool created by Neo4j to facilitate the understanding of problems expressed as graphs. The tool was designed to provide flexibility and resiliency in adapting to changing business models, making it useful for users who want to work with graphs without having to think about writing Cypher or query languages. Bloom is best suited for anyone interacting with a graph, including developers, data scientists, and business users, who can use its point-and-click interface to explore and visualize graphs without needing to write code. The tool has been used in various applications, such as fraud detection and cancer research, and Neo4j plans to continue innovating and adding features to it to showcase the full potential of graph technology.