June 2018 Summaries

The Neo4j community has been actively exploring various graph databases and technologies, including linear regression, the Yen's k-shortest path algorithm, GraphQL for database administration, and more. Alicia Powers, a featured community member, has been part of the Neo4j community for several years and presented popular talks on her work and opinions on the future of graph databases. The community has also developed user-defined procedures to create linear regression models in Neo4j, analyzed the World Cup 2018 Graph with Cypher, GraphQL, and Neo4j Bloom, and explored alternative routes in the California road network using Yen's k-shortest paths algorithm. Additionally, the community has created a new active page cache warmup feature in the Enterprise Edition of Neo4j to preserve the page cache across database restarts. The Neo4j YouTube channel continues to release videos on various topics, including the APOC series and an intro to graph databases.

Tim Ward, CEO of CluedIn, chose Neo4j for its ability to join vast amounts of data together, pattern matching techniques, and path traversals. He finds the platform's flexibility in indexing and scaling to be particularly surprising and valuable. Despite some initial challenges, Ward believes that Neo4j has enabled his company to scale from 15 customers to a database of 280 million nodes with close to a billion relationships, making it an ideal choice for polyglot persistence design. Ward envisions the future of graph technology in his industry as companies adopting graphs as one of the extra types of databases used to solve problems, combined with machine learning techniques, to become data-driven and utilize their data more effectively.

The Neo4j community has been active over the past week, with several new developments and updates. A World Cup Graph dataset has been created, which can be played around with using a hosted Neo4j Cloud instance. A GraphQL API has also been added on top of the database to make it accessible to those not fluent in Cypher queries. The community has also seen a new project called "neo4j-knowledge-graph", an example of a simple, queryable knowledge graph implemented using Neo4j. Additionally, there have been updates and blog posts on various topics such as tuning Cypher queries by understanding cardinality, querying spatial data points, and the introduction of a new Cypher query language. The community is also excited to hear about the acceptance of DATMEAN into the Neo4j startup program, which will allow Bea Hernández to work more closely with Neo4j.

Alicia Powers, a researcher at the New York City Economic Development Corporation, built a recommendation engine using Neo4j to visualize connections in eating and health. She chose Neo4j for its ability to show patterns between different aspects of eating, making it easier to understand data as a visual learner. Alicia found that working with Neo4j brought her closer to her project and the people in the data, allowing her to see personal characteristics such as age, weight, and eating habits. If she were to start over, she would rely more on Neo4j's community resources and support for guidance. Alicia believes the future of graph technology is already here, with applications in various domains, including social networks and neural networks, and expects organizations to increasingly use them effectively.

The Novartis Institutes for BioMedical Research, a pharmaceutical company, has been working on combining heterogeneous data and integrating it into a big knowledge graph to help discover cures for diseases. The underlying problem is how to construct a system of scalable biological knowledge, which involves connecting vast amounts of data and enabling researchers to construct queries for specific triangular relationships between chemical compounds, biological entities, and diseases. Graph database technology plays a significant role in this effort by enabling the capture of the strength of relationships between terms in medical research text, as well as providing a foundation for later queries that link literature to observed chemical or biological data. The company has built a graph database using Neo4j and is now using it to analyze cellular assays and identify connections between compounds, genes, and diseases. The database contains about 30 million nodes, including articles, proteins, gene annotations, and more. By integrating heterogeneous data sources and analyzing relationships, the researchers aim to capture biological knowledge that can be used to understand how compounds interact with cells and develop new medicines.

Graph technology allows companies to leverage existing data stores, such as data lakes and relational databases, to gain insights into their connected data. Creating a graph does not require starting from scratch, but rather building on top of existing relationships between different types of data elements. Graph databases are highly scalable transactional and analytic databases that store data relationships as first-class entities, making it easy to express and persist relationships across many types of data elements. By assembling simple abstractions of nodes and relationships into connected structures, graph databases enable the creation of sophisticated models that map closely to a problem domain. The schema-optional nature of graph databases makes them simple and agile, allowing for easy changes or updates without requiring similar structure for every node. Graph technology has been adopted by companies like LinkedIn and Google, and is now becoming more accessible through mainstream, out-of-the-box graph technology. By using graphs to transfer knowledge of what the organization has done across different departments, companies can get more value from their big data technology and leverage connected data to deliver business insights and actionable results. A case study on Telia, a broadband provider, demonstrates how graph technology can be used to create a smart home platform that simplifies consumers' lives and provides entertainment options, powered by the Neo4j Graph Platform. Overall, graph analytics bring hidden connections in data to light, resulting in lightning-fast queries and numerous use cases across industries.

This week in Neo4j has seen the release of the new temporal and geospatial data types, which are being showcased through a React.js application. The Neo4j ETL Tool has also been featured, with tutorials available on how to use it for data integration. David Meza, Chief Knowledge Architect at NASA, is this week's featured community member, known for his work on building a graph of the public NASA Engineering Network lesson learned database. Jennifer Reif has written a detailed post explaining how to use the Neo4j ETL Tool from the Neo4j Desktop, and Will Lyon has shown how to build an application that makes use of the temporal and geospatial data types in Neo4j 3.4. The Neo4j JDBC Driver 3.3.1 release is also available, with support for Neo4j clusters, routing, read-only transactions, and bookmarks for causal consistency. Additionally, a new dataset called CLEVR graph has been released, which aims to help further research into machine reasoning on graph datasets, while Amass and Mal6raph have been developed as tools for network mapping and malware analysis. The Sinar Project has also built a network chart of the 1Malaysia Development Berhad fund, and there are upcoming events in the world of graph databases.

The Neo4j-JDBC driver has been upgraded to work with recent Neo4j 3.3.x versions and Bolt driver 1.4.6, while work on Neo4j 3.4.x and drivers 1.6.x is in progress. The driver has been improved with the addition of support for in-memory databases, a debug feature, and TrustStrategy configuration options. It now implements the DataSource interface for JNDI lookups and has been updated to use the & separator instead of comma for URL parameters. The documentation has also been updated to explain how to use new features, including a Matlab example. Additionally, the driver can be used with various tools such as Squirrel SQL, Eclipse / BIRT, Jasper Reports, RapidMiner Studio, Pentaho Kettle, and Streamsets. The Neo4j-JDBC driver now supports Causal Clustering, allowing users to route reads and writes to the server with the correct role. It also features a routing context, bookmarks for causal consistency, and support for multiple bootstrap servers. The driver provides an Access Mode (READ, WRITE) feature, which enables transactions to be executed in either read or write mode.

Yaqi Shi, a health informatics student and medicine graduate from China, presents her project "Health-Graph" on using Neo4j to model the connections in the US healthcare system. She was inspired by the complexity of the US system compared to the one in China. Yaqi's process involved researching how different elements in the system relate to each other, developing a data model, and integrating open data sources. Her challenges included dealing with multiple stakeholders, complicated relationships between them, and handling large amounts of data. She used Neo4j to visualize these relationships and demonstrated how her model can be queried to show connections such as money flow between lobbyists and legislators, or medication prescriptions. The presentation aims to provide an idea on how to use Neo4j for industry modeling beyond healthcare.

The text discusses the benefits and applications of "connected data", which refers to the relationships between different data sources, providing a unified view of data that can be used for real-time insights and business decision-making. This approach allows businesses to gain contextual understanding of their data, making operational decisions more informed and effective. The connected data model is particularly useful in industries with complex data landscapes, such as Airbnb's case study, where it helped the company navigate its vast amount of internal data resources and provide transparency to employees. By leveraging graph database technology, organizations can unlock new use cases and create a sustainable competitive advantage in a changing business world.

This week in Neo4j saw the release of the GRANDstack starter kit and Will's screencast on building a full stack application using the framework. The APOC library was used to load JSON APIs into Neo4j, including data from Zendesk and Strava. Py2neo version 4 was also released, providing a client library for working with Neo4j from within Python applications. GraphAware turned 5, and Dirk Mahler, a featured community member, released version 1.4 of his jQAssistant tool. The Lightning Network graph was analyzed, as well as the use of Powershell with Neo4j and building recommendation engines using Recon. Upcoming events include a summer update on Neo4j 3.4 and advanced visualization with Bloom.

The organizers of GraphConnect 2018 are seeking input from attendees and non-presenters alike to help create a list of tips and tricks for presenting at the conference. They want to hear about what makes talks engaging, how past presenters have made experiences memorable or fun, and what speakers could do to improve their presentations. The best ideas will be shared with future GraphConnect speakers and the Neo4j community, and attendees are encouraged to submit their own presentation ideas by July 1st.

Matt Casters, a former Pentaho employee, has joined the Neo4j team as Chief Solutions Architect. He comes from a background of system management and databases, with experience in data integration and the Kettle project, which was open sourced in 2005 and later integrated into Pentaho. At Neo4j, Matt will focus on building solutions integration architecture to accelerate deployment by streamlining integration and making best practices more easily repeatable across different Neo4j projects. He is excited about the growth potential of the technology and its unique graph engine capabilities, which he believes will be a killer combination when integrated with Kettle. Matt can be found online at his blog, GitHub, LinkedIn, and Twitter, and invites community members to get to know him better.

The Py2neo v4 release introduces significant improvements, including a revamped Graph constructor with enhanced settings and authentication options. The API has been expanded to include several logical layers of abstraction, such as the Entity API, which provides CRUD operations on Node and Relationship objects, and the OGM API, which enables creation of GraphObjects that model relationships and properties. Additionally, new features like reporting, analytics, and data science integrations have been added, including support for multiple output formats and integration with popular libraries like numpy, pandas, and sympy. The Cypher lexer has also been implemented for tokenization and syntax highlighting in the interactive console. Furthermore, the `py2neo` command-line tool has been introduced, providing a simple way to manage Neo4j instances, download distributions, and run Cypher queries.

The value of data is finite, but when connected with additional context, its value becomes infinite. This concept is demonstrated through a case study on eBay's use of Neo4j to power an app for Google Assistant, which relies on connected data to provide personalized shopping experiences and improve customer interactions. The article highlights the limitations of traditional relational databases in storing and managing relationships between data elements, and how graph databases like Neo4j can effectively persist these connections, enabling organizations to unlock new insights and business value from their data.

This week in Neo4j saw the release of version 1.2.0 of the Neo4j ETL Tool, which now includes multi-schema support and improved loading processes. The tool makes it easy to convert relational data into a Neo4j graph. Additionally, there are new articles in the Graph Visualization series on using yWorks and Keylines with Neo4j. A video from Adam Hill's PyData London 2018 talk on building a UK corporate interests graph is also available. The community featured Dilyan Damyanov, who has been part of the Neo4j community for a couple of years and has presented his work at the London Neo4j Meetup and Neo4j Online Meetup. Furthermore, there are updates on Graph Gopher, a Neo4j browser for mobile devices, and a new release of the tool that can be used to search for "shady patterns" in corporate ownership data using Neo4j and Cypher queries.

The GraphConnect conference is a global event where graph technology innovators share their stories of success, best practices, and trail-blazing innovation. The call for papers is now open, and the conference will be held on September 20th and 21st at the Marriott Marquis in New York City. Keynote speakers and conference sessions will take place on the 20th, while the 21st will feature Neo4j training classes led by expert engineers. The conference aims to cover a wide range of topics including digital transformation, cloud deployments, artificial intelligence, design thinking, graphs at scale, and integrations. GraphConnect welcomes presentations from developers, software engineers, data scientists, and business executives, regardless of their experience level, offering guidance and support for first-time speakers to help them prepare and build confidence.