July 2020 Summaries

Dr. Lju Lazarevic's presentation at GraphConnect New York City in 2018 delves into tips and tricks for using Neo4j Bloom, a user-friendly data visualization tool, especially beneficial for domain experts without SQL or Cypher experience. Emphasizing the significance of a well-structured data model, Lazarevic explains that a clear representation of domain-specific entities and their relationships enhances the Bloom experience. The talk introduces six generic Bloom patterns—specific path, shortest paths, paths between nodes, exploring depth, commonality, and extended commonality—that facilitate data exploration and visualization without requiring Cypher queries. These patterns help users uncover insights, such as the shortest connections between entities or commonalities in data, by leveraging Bloom's intuitive interface and natural language-like query capabilities. Through examples like Olympic data, Lazarevic illustrates how Bloom's patterns can be used to reveal connections and dependencies within datasets, showcasing the tool's capacity to fill gaps in relationships and provide comprehensive visual insights.

The featured community member this week is Juan Guillermo Gomez, a Google Developer Expert and Neo4j Certified Professional, who has been actively involved in various tech communities and events. He recently published a Neo4j-based Python reference microservice and hosted an Intro to Neo4j tutorial in Spanish. Dr. Alessandro Negro shares his tips for newcomers to graph machine learning and explains common techniques used in recommendations in a new book on Graph-Powered Machine Learning. The Power of Subqueries in Neo4j 4.x has been explored, including the use of correlated sub queries, which were added in Neo4j 4.1. Jasper Blues started a series of blog posts showing how to build a Neo4j-powered mobile game, while Matt Cockayne explains how to pre-populate Neo4j with data using Kubernetes Init Containers and neo4j-admin import. Other topics include building a GraphQL API with Spring Boot, Neo4j, and Kong, and the Proteome Landscape of the Kingdoms of Life graph created by the Matthias Mann Lab.

This week's video showcases a low-code approach to building GraphQL APIs using Neo4j GraphQL Architect, which enables users to infer a schema for an existing database and query it using the in-built GraphiQL IDE. Additionally, Sephi Berry shares his experience with implementing a graph network pipeline using Dagster, while Rik Van Bruggen revisits his previous work on the network of executives of Belgian public companies. Preet Kanwar continues his blog post series on building a Neo4j backed application with Spring Boot and Kong, an open-source API gateway and microservice management layer. The community also features Demian Bellumio, Global Vice President of Augmented Intelligence at NEORIS, who shares his experience using graphs for machine learning and presents on the Neoris COVID-19 HealthCheck Graph.

The Graph Data Science (GDS) library has been released in version 1.3, which leverages Neo4j 4.0.1 and 4.1 and includes enterprise features such as role-based access control and multi-database support. The library now offers over 50 unsupervised algorithms, including a preview of graph embeddings, which are learned through neural network models or linear algebra. Performance has been significantly improved, with production-quality algorithms being up to 85% faster than the deprecated Graph Algorithms library. The GDS framework provides a production-ready platform for data science at an enterprise scale, with features such as premium support, three tiers of algorithms, flexibility and expressivity, data science-focused features, and compatibility with the Neo4j product portfolio.

This week, the Neo4j community has been busy with various projects and announcements. Adam Cowley released a video on building web applications with Neo4j and TypeScript, while Stefan Dreverman continued his series on building low-code platforms with Neo4j. Alicia Frame announced the release of version 1.3 of the Graph Data Science Library, which includes graph embedding algorithms and support for weights in node similarity. Preet Kanwar started a blog post series on building a Neo4j backed application with Spring Boot and Kong. Meanwhile, Paul Jongsma, a featured community member, has been actively contributing to the Neo4j community, sharing his graph journey through interviews and podcasts. The community has also seen various tutorials and explanations on topics such as loading CSV data into Neo4j, using APOC library, and visualizing graphs with 3d-force.

The text discusses the implementation of an enterprise knowledge graph using Google Drive and Neo4j. The knowledge graph is designed to connect discrete pieces of information together with context, providing fast and flexible querying capabilities. It leverages AI/ML-based predictive capabilities to provide relevant suggestions and answers in real-time. The system extracts metadata from Google Drive documents and ingests them into Neo4j, creating a graph that can be queried and analyzed. The knowledge graph is extended by adding n-grams, which are used to calculate document similarity and identify key terms. The system provides insights into the connectedness of documents, popularity of certain words/phrases, and similarity between documents based on n-gram overlap and position in the hierarchy.

Hi graph gang, In this week's video, Will Lyon introduces GraphQL Architect, a Neo4j Graph App for building and deploying GraphQL APIs. Stefan Dreverman builds the UI-engine in the low-code platform, Rik Van Bruggen explains what Recommender Systems and Contact Tracing have in common, and Mahjoub Saifeddine builds a basic web crawler using Go and Neo4j. Josh Thurston has written a tutorial showing how to use Neo4j with Python's popular Flask web application framework. This week's featured community member is Koji Annoura, Chief Technology Officer at UTI, Inc., who participates in almost everything in the Neo4j community, including hosting events and helping fellow members. The latest video is part of a series by Will Lyon showing how to build a GRANDstack Real Estate Search App, where he introduces GraphQL Architect for building and deploying GraphQL APIs. In other news, Preet Kanwar started a blog post series on building a GraphQL API with Spring Boot, Neo4j, and Kong, while Davey Brown shared idenavoice, a solution to online voting using the GRANDstack. João Esperancinha shows how to handle data related to investigator cases using the GRANDstack. A hackathon called The Effects of the Seasons uses spaCy library to extract sentence structure from COVID-19 documents and analyse results in Neo4j. Stefan Dreverman continues his series on building a low-code platform with Neo4j, while Mahjoub Saifeddine builds a basic web crawler using Go and Neo4j. Alicia Frame announced the preview release of version 1.3 of the Graph Data Science Library, which includes graph embeddings, and Gerrit Meier released SDN/RX 1.1.1 with improvements and bug fixes. Josh Thurston wrote a tutorial on Flask User Auth with Neo4j, while Paula González shared her favorite tweet about the Connections webinar.

Fabric is a new feature in Neo4j 4.0 that allows storing and retrieving data in multiple databases using a single Cypher query, providing infrastructure and tooling for data federation and sharding. It's an Enterprise-only feature, not available for the Community Edition. To set up Fabric, users need to download Neo4j Desktop, create a new database, and configure the fabric.database.name setting. Once configured, users can use Fabric to query single graphs or multiple graphs by specifying the database name in the query. Fabric also supports creating data, updating data, and running correlated subqueries. The feature provides an infrastructure for separating data into different geographical locations, making it easier to control access to sensitive information.

Neo4j is a graph database that stores data as nodes and relationships, providing a flexible and schema-optional way to model complex relationships between entities. Unlike traditional SQL databases, which rely on tables and joins, Neo4j uses an array-based structure for queries, allowing for faster performance in certain use cases. The key to understanding Neo4j lies in reimagining data as a graph, using nodes and relationship types to represent complex connections between entities. This approach can lead to more efficient querying and modeling of data, but also requires creative thinking and experimentation to achieve optimal results. By leveraging the properties of relationships and node structures, users can build complex queries and models that take advantage of Neo4j's unique strengths.