June 2022 Summaries

The Neo4j Code Golf contest is a gamified, leaderboard-based programming competition where developers showcase their Cypher coding skills to win prizes. The contest features three different problems that must be solved using Neo4j, a graph database, and is open to everyone regardless of experience level. Participants submit short Cypher queries with low database hits to win the contest, with the winner receiving $4000 in total across all levels. The contest provides access to learning resources, including a GraphAcademy Course on Cypher Fundamentals, as well as experts within Neo4j via Discord and the community. Submissions will be accepted starting June 30, 2022, until September 15, 2022.

Neo4j has launched a new product called Neo4j Ops Manager, which helps manage all Neo4j deployments at scale from a single user interface. The tool allows teams to monitor and administer Neo4j databases, hosts, and instances on premises or in the cloud, regardless of their size or complexity. With Neo4j Ops Manager, operations teams can easily manage multiple Neo4j deployments, gain actionable insights into database performance, and optimize resource utilization rates. This new product enhances Neo4j's value proposition for enterprises, particularly in terms of turnkey monitoring, centralized management, and actionable analytics. By using Neo4j Ops Manager, teams can improve their productivity, reliability, and performance, making it easier to manage large-scale graph databases like those used in AI/ML projects.

The modern data stack is composed of various technologies, including databases, data warehouses, and data lakes, which are interconnected with powerful ETL tools. Neo4j aims to be a part of this landscape by providing connectors for efficient import and export of data. The Neo4j Kafka connector enables bi-directional data flow between Neo4j and Apache Kafka, while the Spark connector allows batch processing between Neo4j and Apache Spark data frames. Additionally, the Data Warehouse Connector facilitates movement of data between Neo4j and popular data warehouses like Snowflake and BigQuery, and the BI Connector provides a read-only interface for visualizing or combining data from multiple databases in business intelligence tools like Tableau.

The Neo4j Ops Manager is a tool designed to simplify the monitoring, administration, and operation of Neo4j Enterprise databases, providing an enjoyable experience for administrators. It manages Neo4j instances regardless of type, deployment, or location, and offers features such as estate view, asynchronous management, four pillars (monitoring, administration, operations, and integration), secure and protected access, loose-coupled architecture, lightweight and efficient user interface, extensible and explainable design, and a three-tier architecture. The tool provides initial features in monitoring, administration, and operations, including support for single instances and clusters, agent management, password management, DBMS rename, and supporting frameworks. Neo4j Ops Manager is available as an official product and can be evaluated with a trial license of Enterprise Edition Neo4j database.

The text highlights a project by Kacha Mukabe, a grand winner of the 2021 Leonhard Euler Idea Contest, which uses Neo4j graphs to create a developer community platform called Developers.Zed. The platform aims to make the local developer community more visible and connect developers based on their relationships and skills. Despite having no graph database background, Kacha was able to build the application in just two weeks using GRANDstack and the Neo4j Cypher language. He faced challenges with reconciling GraphQL queries and working with React, but overall found it easier than expected. The source code for the front-end and backend are open on GitHub, allowing others to learn from his thought process and apply similar schematics to their own projects.

For developing a Neo4j Data Access for .NET Core, it is recommended to set up a free Neo4j Aura instance or use Neo4j Desktop on your local machine. The Neo4j Official Driver NuGet Package should be installed in the data access project. The app settings file should contain the Neo4j Bolt/Neo4j connection string, username, password, and database instance name, with additional security measures implemented to protect the password. The Neo4j Data Access class is injected as a scoped dependency, and the IAsyncDisposable interface provides a mechanism for releasing unmanaged resources asynchronously. The wrapper methods manage the Neo4j session using IAsyncDisposable, allowing for easy implementation and reusability. Domain repository methods can be used to prepare Cypher queries and parameters, utilizing map projections to construct data from nodes, relationships, and other values. This approach helps avoid extra boxing and unboxing, making it easier for clients to parse the JSON to their desired object types.

The latest update of Bloom 2.3 now includes Graph Data Science features, allowing users to select from various algorithms and utilize the GDS plugin or AuraDS on self-managed databases. This development enables experimentation and visualization of supply chain data using Neo4j's Graph Data Science and Bloom, as demonstrated by Zach Blumenfeld in his blog series. Additionally, local GraphSummits are being held in EMEA and APAC regions, providing opportunities for attendees to engage with the community and learn about the capabilities of graph technology. Various tutorials and projects have been created to help developers get started with Neo4j and graph databases, including a Java-based AWS Lambda application and a .NET Core C# microservice. The use of graph databases is also highlighted as a solution to problems that other databases cannot address, such as surface critical information and analyzing semantic structure.

The latest update to Graph Data Science (GDS) introduces several significant features to make graph data science easier to use. The new algorithms include K-means clustering and Leiden for community detection, which provide more accurate and efficient results. Additionally, the update includes autotuning for machine learning pipelines, source and target node filtering for KNN and Node Similarity, and arrow support for fast graph projection, database creation, and graph export. These features significantly improve performance, allowing users to train models up to 10 times faster than before. The Graph Data Science Python Client also receives visual progress logging, which provides a better user experience. Furthermore, the update introduces new alpha-tier algorithms, including Leiden and K-means clustering, as well as node regression pipelines for predicting numerical property values. The integration of Apache Arrow enables users to import and export massive graphs directly into GDS at speeds up to 30 million objects/second. Overall, this update aims to make graph data science more accessible and user-friendly, empowering data scientists to work with connected data more efficiently.

Neo4j Bloom 2.3 is now available, offering a powerful combination of Graph Data Science and data visualization capabilities. Users can leverage rule-based styling to easily visualize results from algorithms such as centrality measures or community algorithms. This latest release brings together the power of Neo4j Graph Data Science Library with AuraDS users and those with Neo4j Graph Data Science setup on their databases. Existing users will receive the update automatically, while others can install it using the Graph Apps drawer or download updates for server plugin users or self-hosted applications. The update also includes features tailored to specific user needs, such as a supply chain demo showcasing how to leverage Bloom with the latest Graph Data Science integration. Users are invited to share their experiences and provide feedback to help design an experience tailored to their needs.

Graph data science is being applied to analyze a logistics network using Neo4j Graph Data Science (GDS) and Bloom. The goal is to gain insights into operational load, flow control, and regional patterns in the supply chain. The analysis starts with visualizing the raw data in its tabular form, which is difficult to parse, but becomes more transparent when ingested into a graph model. A graph data model is created to represent nodes and relationships, including transportation services, airports, and stages. The Louvain algorithm is used to identify regional interdependence within the network, and community detection reveals distinct regions or "communities" of interconnected stages. Degree centrality measures operational load, while betweenness centrality evaluates flow control and bottleneck risk. These algorithms are applied using Neo4j GDS and Bloom, providing an intuitive no-code interface for analysis. The study demonstrates that graph data science can be used to analyze complex supply chain networks without relying on geographical identifiers.

The author reflects on the impact of Harvey Milk's legacy, specifically his "Hope" speech, which emphasized the need for hope in the face of adversity. The author notes that 30 years have passed since Milk's death and has seen significant progress in LGBTQ+ rights, including increased visibility, acceptance, and legal protections. Despite concerns about potential backlash, the author maintains hope that the arc of history will continue to curve toward justice. They attribute this hope to the growing awareness and representation of LGBTQ+ individuals, as well as the decline of discriminatory attitudes over time. The author concludes by expressing their commitment to leaving readers with hope, echoing Milk's words.

This Week in Neo4j highlights various updates and news from the Neo4j community, including a partnership between Neo4j AuraDS and Google Cloud Vertex AI to build graph-based machine learning models. The platform allows developers to easily use graph structure to compute algorithms or create embeddings, increasing accuracy and reliability in machine learning pipelines. Additionally, there are resources available for building GraphQL APIs with managed cloud services, a featured community member profile, a Cypher cheatsheet, GraphConnect recordings, Centrality Algorithms tutorials, and updates on Neo4j Migrations and Graphlytic releases.

Normalization in traditional relational databases serves a specific purpose, primarily designed to protect data integrity by reducing redundancy. However, graph computing models this concept differently, focusing on relationships as the fundamental structure of the database. Graph databases like Neo4j achieve efficiency without sacrificing data integrity through their inherent design, which doesn't require denormalization or reassembly of data for relationships. This approach enables query efficiency and parallel execution across graphs, allowing for federated queries, horizontal scaling, and hybrid cloud queries. The traditional relational model's emphasis on normalization is seen as less practical at scale, where graph models offer a more natural fit for complex, interconnected data. By building databases with graph structures from the outset, organizations can avoid unnecessary denormalization and re-normalization efforts, streamlining their data management processes.

We are thrilled to announce that the Neo4j GraphSummits are hitting the road, traveling across Europe and Asia-Pacific with stops in 17 cities. Our experts will share the latest trends for innovating with graphs, provide insights on data challenges, and showcase new product releases while networking opportunities abound. You can expect to meet our team, learn from customers' success stories, participate in hands-on workshops, and stay updated on the latest developments in graph technology, all in a fun and social setting.

Neo4j has released a new version of its Graph Data Science Platform, making it easy for data scientists to start using Google Cloud with Neo4j AuraDS, a fully managed graph data science offering. This platform allows users to amplify their machine learning pipelines by feeding knowledge of the graph into ML models to enhance predictability. Graphs are becoming increasingly important in various industries, such as pharmaceuticals and finance, where they help analyze relationships between people, places, objects, and events. Neo4j AuraDS provides access to over 65 state-of-the-art graph and machine learning algorithms, along with a simple hourly pricing model, making it easy for users to understand and manage costs associated with compute. The platform is designed to be user-friendly, allowing users to create instances with minimal configuration, resize them as needed, and pause them when not in use. AuraDS can be integrated with Vertex AI, providing a seamless workflow for machine learning pipeline development and deployment. The integration allows users to generate graph embeddings using FastRP, train models on them, and deploy the trained model for new predictions. With Neo4j Graph Data Science, data scientists can focus on building their business while leveraging a complete platform that makes it easy to start and take their use case from hypothesis to production without requiring a team of software engineers or database managers.

The Neo4j community has been abuzz with exciting news and announcements from GraphConnect 2022 in Austin, Texas. The event saw the introduction of new tools and features, including Neo4j Workspace in UserTools, Neo4j Data Warehouse Connector, Neo4j Operations Manager, and Neo4j Cluster and Fabric. Additionally, Neo4j GDS introduced Apache Arrow integration for graph projections, as well as machine learning algorithms like K-means clustering and node regression ML pipeline. The community is also celebrating the launch of Alessandro Negro's new book "Knowledge Graphs Applied", which guides readers in creating knowledge graphs from first principles to developing advisor applications. Furthermore, users can now utilize the ImageToGraph tool to create graph representations of images and explore various data science tools like the Neo4j Graph Data Science Python Client. To stay organized, format_cypher.java is available to clean and organize Cypher queries. The community is also being encouraged to retweet a tweet from @Ifeanyidiaye showcasing their love for Neo4j and graph databases.

I've discovered fika as part of Neo4j's Swedish culture in London. Fika is a moment to connect with others, often over coffee and pastries, sharing the experience rather than just enjoying individual components. It's not about the food or atmosphere alone but about creating connections and being present with someone else. I've explored various fika spots in London, including Hej Coffee, Ole & Steen, Maître Choux, Cedric Grolet, and NoNo, highlighting their unique offerings and experiences. These locations showcase the essence of fika beyond traditional Scandinavian settings, making it accessible to everyone. Ultimately, fika encourages taking a break and connecting with others in our busy lives.

The Neo4j team has hosted several digital events, including Connections, webinars, and live demos, with a focus on showcasing the capabilities of their graph database platform. These events have covered various topics such as Graphs4Good, product updates like Neo4j 4.4 and Bloom 2.0, and use cases like fraud detection, supply chain optimization, and knowledge graphs for skills discovery. The live demo series features eight different topics, each offering a 30-minute introduction to the graph data platform and its applications in various industries.

Neo4j has been hosting various digital events including Connections, Webinars, and Live Demos for the past year. The company's Global Marketing Programs team has hosted 17 Connections events, 51 webinars, and almost 100 live demos. These events cover a range of topics such as Graphs4Good, Neo4j 4.4, Neo4j Bloom 2.0, Graph Data Science 2.0, and various use cases like Supply Chain, Fraud Detection, Knowledge Graphs, Graph Data Science & Machine Learning, and more. The live demos offer a close-up look at Neo4j's capabilities and are scheduled for every Wednesday and Friday at 10:00 a.m. PT / 1:00 p.m. ET.

Neo4j has introduced a unified tools experience called Workspace, which brings together Data Importer, Bloom, and Browser to help users get up and running with graph databases quickly and efficiently. This new feature integrates Data Importer, enabling users to import multiple datasets, sketch their data model, map categories, import data, and explore it in their Neo4j graph database. Graph Data Science has also been enhanced with new features such as Apache Arrow Integration for Graph Projections, allowing users to import and export massive graphs up to 8 million objects/second, and new algorithms like Leiden and K-means clustering. Additionally, Neo4j Operations Manager has been introduced, providing a graphical UI-based tool for monitoring and administering all Neo4j databases, instances, and clusters from a central interface. New clustering changes in Neo4j 5.0 offer horizontal scale without requiring existing servers to be scaled vertically, making operational issues associated with managing large graphs easier to manage. The new Neo4j Data Warehouse Connector provides a simple way to move data between Neo4j and popular Data Warehouses like Snowflake, BigQuery, Amazon Redshift, and Microsoft Azure Synapse Analytics.

The DevZone at a developer conference offers a unique experience with comfortable sofas, t-shirt printing, and a sketch artist. It's also set up for technical deep dives with demo stations and experts available to discuss challenging issues. The space includes a scavenger hunt with the chance to win drones through a raffle, as well as a video booth where attendees can share their thoughts and participate in fireside chat videos.

This Week in Neo4j: Community Announcement, Clinical Trials ML, Graph Embeddings, Real-Time Analytics, Graph Neural Networks, and More`: The Neo4j community platform has been launched with a new community platform hosting more content and scaling to infinite conversations. A recent study used Neo4j to optimize the pharmaceutical research workflow by estimating the probability of a drug receiving approval in clinical trials. Machine learning approaches have also been applied to graph databases for various applications such as predicting clinical trial outcomes, building GraphQL APIs, and analyzing biomedical information. Graph embeddings have been explored using the node2vec approach, while knowledge graphs are being used to map and analyze biomedical information. Various Neo4j experts and developers are sharing their insights on topics like natural language processing, path-finding algorithms, and the impact of knowledge graphs. The goal of the community remains the success of its members, helping them grow their relationships in the world of graphs.

The Neo4j Graph Data Science Python client is a new tool designed to help users integrate Neo4j graph algorithms into their Python analytical workflows. The client's syntax mimics the GDS Cypher procedures, allowing users to project and execute graph algorithms using pure Python code. With the release of version 2.0 of Neo4j Graph Data Science, a supporting Python client has been introduced, enabling seamless integration with popular data science libraries such as Pandas. The client can be installed using pip and provides methods for executing Cypher statements, projecting in-memory graphs, running graph algorithms, and more. Users can leverage the client's benefits, including automatic conversion of algorithm outputs into Pandas DataFrames, to simplify their workflow and enhance their analytical capabilities.

Cybersecurity is crucial to protect against rising cyberattacks, which can be mitigated with advanced data solutions that enable real-time correlation and analysis of connections at scale. Effective security posture involves monitoring assets continuously and automatically, identifying vulnerabilities, and implementing defense in depth strategies such as zero-trust approaches. Budgeting for cybersecurity requires understanding the likelihood and potential cost of breaches to present proposals with a projected return on investment. Training employees is essential to reduce human error-related breaches, while policies and procedures like least privilege, patching, authentication, backups, and vendor security evaluations are critical to minimize risks. A graph-based approach can enable organizations to analyze their security posture, identify vulnerabilities, and automate monitoring and alerts.

A true, native graph database can help solve complex problems by visualizing relationships between data entities. The right graph, which represents objects and relationships in a simple yet effective way, is essential for finding solutions to real-world problems. Graph databases like Neo4j can be used to model data in a way that was not possible with traditional relational databases, enabling the use of algorithms from other fields to solve complex problems such as fraud detection, money laundering, and supply chain management. By using graph databases, businesses can simplify their data, find solutions, and scale their models as needed, without having to manually transform or remodel their data. With Neo4j's AuraDB, a fully managed and self-maintained graph database system in the cloud, businesses can start solving real-world problems today, for free.